Skip to main content

Full text of "Elucidation of imprinting mechanisms and phenotypes in Prader-Willi syndrome mice"

ELUCIDATION OF IMPRINTING MECHANISMS AND PHENOTYPES IN 
PRADER-WILLI SYNDROME MICE 



By 

STORMY JO CHAMBERLAIN 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE 
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE 
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 



2003 



ACKNOWLEDGMENTS 

The author thanks the past and present members of the Brannan lab~Winthropian 
Peery, Karen Johnstone, Chris Futtner, Jessica Walrath, Susan Blaydes, Tom Simon, 
Mike Elmore, Todd Adamson, and Missy Shelley--for making lab such a fun place to 
work, for tolerating her daily, and for sticking together through some tough times. She 
extends special gratitude to Karen Johnstone for her gracious help with this document. 

She also thanks Dan Driscoll for being the butt of her jokes and for not letting her 
forget that she, too, can be a butt, and Jim Resnick for accepting responsibility for her 
and for keeping all of us together. 

The author extends her deepest gratitude to Cami Brannan for allowing her to 
work on this project, for her ideas, for giving her someone to look up to, and for years too 
few of friendship. 

Finally, the author thanks her parents for all of their support and for making her 
come home when she needs to. 



ii 



TABLE OF CONTENTS 



ACKNOWLEDGMENTS ii 

LIST OF ABBREVIATIONS vi 

ABSTRACT viii 

CHAPTERS 

1. INTRODUCTION 1 

Genomic Imprinting 1 

The Prader Willi and Angelman Syndromes 4 

2. IMPRINTED REGULATION OF UBE3A 9 

Introduction 9 

Materials and Methods 11 

Strains and Matings 1 1 

Identification of Polymorphisms 12 

RT-PCR 12 

Results 13 

Discussion 15 

3. IMPRINTED TRANSGENES 23 

Introduction 23 

Use of Transgenic Mice to Study Gene Regulation 23 

Materials and Methods 26 

Screening the BAC Libraries 26 

Probe Preparation 28 

Verification of BAC Clones 29 

End Sequencing BAC Clones 29 

Pulse-Field Gel Electrophoresis 30 

Southern Blot 30 

Transgenic Mice Production 31 

Mouse Husbandry 32 

Mouse Strains 32 

Results 32 

YAC Studies 32 

129/Sv BAC Library Screen 33 

iii 



RecA Mediated Homologous Recombination 34 

RecET Mediated Homologous Recombination 36 

C57BL/6J BAC Library Screen 36 

Lambda Red Mediated Homologous Recombination 38 

Production Of Transgenic Mice 39 

Characterization Of Transgenic Lines 40 

The Marked Endogenous Snrpn Locus 41 

Imprinted Expression From Transgenes 42 

Discussion 46 

4. STRAIN-DEPENDENT DIFFERENCES IN PHENOTYPE 64 

Introduction 64 

Materials and Methods 66 

Strains and Matings 66 

Culling and Fostering 67 

Identification of Polymorphisms 67 

RT-PCR 68 

Northern Blot Analysis 69 

Results 70 

Discussion 72 

5. TRANSGENIC RESCUE OF THE PWS-IC DELETION MOUSE 80 

Introduction 80 

Materials and Methods 85 

Screening the BAC Libraries 85 

Probe Preparation 86 

Verification of BAC Clones 86 

End Sequencing BAC Clones 87 

Pulse-Field Gel Electrophoresis 87 

Southern Blot 88 

Transgenic Mice Production 88 

Mouse Husbandry 89 

Mouse Strains 89 

Results 90 

Screening the BAC Library for Clones 90 

Production and Characterization of Transgenic Mice 91 

Transgenic Rescue Experiments: 454 Transgenic Lines 94 

Transgenic Rescue Experiments: 380 and 1707 Transgenic Lines 95 

Discussion 96 

BAC Isolation 96 

Rearranged Transgenes 97 

Transgenic Rescue 99 

6. CONCLUSIONS AND FUTURE DIRECTIONS 1 10 



iv 



REFERENCES 114 

BIOGRAPHICAL SKETCH 121 

v 



LIST OF ABBREVIATIONS 



AS 


Angelman syndrome 


AS-IC 


Angelman syndrome imprinting center 


AS-SRO 


Angelman syndrome smallest region of deletion overlap 


BAC 


bacterial artificial chromosome 


cDNA 


complementary deoxyribonucleic acid 


CpG 


cytosine-phosphate-guanine 


DNA 


deoxyribonucleic acid 


DNasel 


deoxyribonuclease I 


IAP 


interstitial A particle 


IC 


imprinting center 


ICR 


imprinting control region 


Igft 


insulin-like growth factor-2 gene 


Igf2r 


insulin-like growth factor-2 receptor gene 


Ipw 


imprinted in Prader-Willi gene 


kb 


kilobase 


Magel2 


mage-like-2 gene 


Mb 


megabase 


Mkrn3 


makorin-ring-3 gene 


mRNA 


messenger ribonucleic acid 


NCBI 


National Center for Biotechnology Information 



vi 



NCI 


National Cancer Institute 


Ndn 


TLT J* 

Necdin gene 


PCR 


polymerase chain reaction 


PWS 


Prader-Willi syndrome 


PWS-IC 


Prader-Willi syndrome impnnting center 


PWS-SRO 


Prader-Willi syndrome smallest region or deletion overlap 


RNA 


ribonucleic acid 


RNasel 


ribonuclease I 


RPCI 


Roswell Park Cancer Institute 


RT 


reverse transcription 


RT-PCR 


reverse transcription polymerase chain reaction 


SmN 


small nuclear ribonucleoprotein particle N protein 


T"» XT A 

snoRNA 


small nucleolar nbonucleic acid 


Snrpn 


small nuclear ribonucleoprotein particle N gene 


Snurf 


Snrpn upstream reading frame 


Ube3A 


ubiquitin 3 A ligase 


UPD 


uniparental disomy 


YAC 


yeast artificial chromosome 


ZNF127 


human zinc finger protein 127 


Zfpl27 


murine zinc finger protein 127 



vii 



Abstract of Dissertation Presented to the Graduate School 
of the University of Florida in Partial Fulfillment of the 
Requirements for the Degree of Doctor of Philosophy 

ELUCIDATION OF IMPRINTING MECHANISMS AND PHENOTYPES IN 
PRADER-WILLI SYNDROME MICE 

By 

Stormy Jo Chamberlain 
May, 2003 

Chair: James Resnick 

Major Department: Molecular Genetics and Microbiology 

Genomic imprinting is an epigenetic phenomenon in which genes are expressed 
exclusively from one parental allele or the other. Prader-Willi syndrome (PWS) and 
Angelman syndrome (AS) are human genetic disorders resulting from the absence of 
genes that are subject to genomic imprinting. PWS is a contiguous gene syndrome that 
results from the absence of two or more paternally expressed genes, while AS is a 
disorder that can result from the absence of a single maternally expressed gene, UBE3A, 
which lies in the same region. 

This study uses a mouse model for PWS in which paternal transmission of a PWS 
imprinting center (PWS-IC) mutation results in complete loss of local paternal gene 
expression. This targeted mutation accurately models both the molecular and phenotypic 
aspects of PWS, and it has been applied to both of these aspects in this study. 



viii 



It has first been shown that imprinted expression of Ube3a, most likely results 
from the regulation of a paternally expressed transcript that is antisense to the Ube3a 
gene, suggesting an interesting model for the regulation of gene expression in this 
imprinted region. The minimum sequence that is sufficient to confer proper imprinted 
expression of the paternally expressed genes to transgenes in vivo was then identified. 
This provides a stringent assay that can be used to test the proposed model of imprinted 
gene expression. 

Transgenic rescue of the PWS mouse was also used in our mouse model to 
correlate specific attributes of the PWS phenotype to specific gene products. A transgene 
that shows expression of a single small nucleolar RNA (snoRNA) cluster rescues the 
small stature phenotype of PWS mice, while three other genes were found to have no 
obvious effect on the PWS phenotype in mice. We have also found that the PWS 
phenotype is less severe on certain strain backgrounds, suggesting the presence of 
modifier genes that may ameliorate the PWS phenotype. These modifier genes may help 
to identify gene pathways involved in PWS as well as future targets for gene therapy that 
may ultimately help the PWS patient. 



ix 



CHAPTER 1 
INTRODUCTION 

Prader-Willi and Angelman syndromes (PWS and AS, respectively) are two 
human genetic disorders that result from deficiencies in genes that are subject to genomic 
imprinting. The causative genetic regions for both of these syndromes have been mapped 
to approximately 3 Mb in the 15ql l-ql3 region. They provide a clinically and 
scientifically interesting situation in which to study the mechanisms and phenotypes 
involved with one region of genomic imprinting. 

Genomic Imprinting 

Genomic imprinting is a phenomenon in which genes are expressed in a parent- 
of-origin specific manner. 1 Both parental genomes, in fact, are absolutely required for 
normal development of a mammal (mouse). Work from the laboratories of Davor Solter 
and Azim Surani showed that embryos with two genetic contributions from the same 
parent would not develop normally. 2,3 Embryos with two male pronuclei (androgenetic) 
developed into tissues that were mostly extraembryonic in nature, while those with two 
female pronuclei (parthenogenetic) developed into tissues that represented the embryo 
proper. 2 Cattanach showed that uniparental disomy, or the inheritance of specific 
chromosome segments from a single parent, also led to developmental abnormalities, 
further suggesting the importance of both parental genomes. 4 This also led to the idea 
that it may be specific genes or regions that were differentially regulated between 
maternal and paternal alleles, and thus responsible for the failure of parthenogenetic or 



1 



androgenetic embryos to develop correctly. These experiments suggested that certain 
genes in the mammalian genome were subjected to genomic imprinting and are expressed 
exclusively from the paternally inherited allele and silenced from the maternally inherited 
allele, while others are expressed from the maternally inherited allele and repressed from 
the paternally inherited allele. 5 

The parent-of-origin specific expression is maintained from generation to 
generation, and so with each passage through the germline, the past parental "label" of an 
imprinted gene is erased, and the gene is "re-labeled", based on the parent that it is now 
inherited from. 6,7 The ability of the two parental alleles to always be distinguished from 
one another is fascinating, especially since these alleles can be identical. The existence 
of identical alleles that are somehow regulated differentially indicates that genomic 
imprinting involves an epigenetic phenomenon. 8 Epigenetic, meaning 'outside 
conventional genetics', 9 refers to the distinguishing mark being something other than 
changes in sequence. There is now considerable evidence to suggest that chemical 
modifications, such as methylation or acetylation, to DNA and proteins associated with 
DNA are responsible, in part, for the epigenetic changes that distinguish one parental 
allele from another. 10 DNA methylation has been known to be associated with silenced 
alleles, although some preferentially expressed alleles are known to be associated with 
DNA methylation. 61113 The histone code has also been shown to be associated with 
either the maintenance or establishment of imprinted domains. The current 
understanding of how DNA methylation collaborates with proteins that modify 
nucleosomes to establish a stable chromatin state and epigenetically regulate gene 



3 

expression is nicely reviewed by Rudolf Jaenish and Adrian Bird in the March 2003 
volume of Nature Genetics. 9 

Is extremely important to understand the cis acting regulatory sequences that 
harbor the epigenetic changes and translate them to changes in gene expression, since 
gene expression from one parental allele or the other is the ultimate goal of genomic 
imprinting. While many imprinted regions, such as H19-Igf2 and Igf2r have been studied 
in great depth, it has become apparent that different imprinted regions are not necessarily 
subjected to the same regulation paradigms. 

The PWS/AS region of genomic imprinting is an attractive imprinted region to 
study for several reasons. First, the imprinted region spans a distance of approximately 3 
Mb. 14 The mechanism of a control element that influences gene regulation at such great 
distances is particularly intriguing. Secondly, this region seems relatively simple. At the 
outset of this study, there were four known paternally expressed genes and one known 
maternally expressed gene in the region. 1519 It was also known that this region was 
regulated by a bipartite imprinting center, with one portion regulating the maternally 
expressed gene, and the other regulating the paternally expressed genes. 20 Thirdly, at 
least two mouse models with disrupted imprinting in the PWS/AS region suggested that 
the imprinting mechanism was conserved between mouse and human, providing a genetic 
system in which to study imprinting. 2122 Most importantly, the PWS/AS region has 
clinical significance, as it was originally identified as a genetic region that was 
responsible for human disease. 



The Prader Willi and Angelman Syndromes 

PWS is characterized by neonatal hypotonia, failure to thrive, hypogonadism, 
cryptorchidism, and poor suckle. 23 By 18-36 months, the failure to thrive and poor 
suckle symptoms subside, and PWS children begin to develop obesity that is complicated 
by hyperphagia, physical and mental developmental delay that leads to moderate mental 
retardation and small stature, and behavioral problems reminiscent of obsessive- 
compulsive disorder. 23 AS is characterized by severe mental retardation, ataxic gait, 
inappropriate laughter, happy affect, and almost absent speech. 24 Each of these disorders 
occur at a frequency of approximately 1 in 15,000 live births. 24-26 While PWS and AS 
are relatively rare, they provide a closed genetic system in which to investigate human 
disorders, such as obesity, failure to thrive, ataxia, and mental retardation, that are much 
more common. 

Ledbetter et al. first noticed that patients with PWS often had cytologically visible 
deletions of human 15ql l-ql3. 14 Butler and Palmer later reported that these same 
deletions were occurring on the paternally inherited chromosome. 27 Finally, Nicholls et 
al. determined that maternal uniparental disomy was also a mechanism by which PWS 
could occur. 28 This led to the suspicion that PWS occurs as from the physical or 
functional deletion of the genes located on the paternally inherited chromosome. AS, on 
the other hand was known to be associated with similar deletions that were occurring on 
the maternally inherited chromosome. 29,30 Paternal uniparental disomy was also shown 
to result in AS. 31,32 AS was found to be the result of the physical or functional deletion 
of maternal 15qll-ql3. 



5 



Patients with PWS lack expression of all of the paternally expressed genes in the 
region and can be divided into 3 genetic classes— maternal uniparental disomy patients, 2 
deletion patients, 14 and imprinting defect patients. 6-20 Maternal uniparental disomy 
patients have two chromosomes 15 that were inherited from their mother, presumably as 
a result of nondisjunction, followed by reduction in the zygote. Absence of a paternal 
allele of 15ql l-ql3 causes the absence of expression of the paternally expressed genes. 
The deletion patients have a 3-4 Mb deletion of the paternally inherited 15ql l-ql3, 
which is facilitated by repeated regions that flank the entire imprinted domain. These 
repeats cause this type of deletion to occur at a comparatively high frequency. Patients 
with this deletion are physically missing the paternally inherited allele from which genes 
are expressed. Imprinting mutations were identified in patients that had inherited both a 
maternal and paternal chromosome 15ql l-ql3, but the paternal chromosome had a 
maternal epigenotype as assayed by methylation at the SNRPN gene. Some of these 
imprinting mutations were found to be due to deletions of the PWS-IC and result in the 
silencing of the paternally expressed genes. 61 1 ' 13 - 20 Other imprinting mutations show no 
apparent deletion or mutation in the PWS-IC, and are thought to be epimutations of 
unknown origin. 33 

AS patients can be divided into five genetic classes— paternal uniparental disomy 
patients, 31,32 deletion patients, 29,30 imprinting defect patients, 6,20 UBE3A mutation 
patients, 34,35 and biparental patients with an unidentified genetic lesion. Paternal 
uniparental disomy patients have two chromosomes 15 from their father. Again, 
nondisjunction is thought to be responsible for this, but it occurs less frequently in males 
than females, accounting for the reduced frequency of paternal uniparental disomy 



patients. AS deletion patients are physically missing the maternal allele of 15ql l-ql3, 
and thus the two maternally expressed genes in the region. Imprinting mutations that 
result from microdeletion of the AS-IC disrupt the maternal pattern of gene expression. 
These imprinting mutations were shown to be small deletions upstream of the PWS-IC 
deletion mutations and never overlap them. Imprinting mutations in AS patients have 
been identified by mutation of the putative AS-IC and not by epimutation, since no 
known methylation mark has been associated with AS (A. Lossie and D.J. Driscoll, 
personal communication). Finally, the two biparental classes of AS patients indicate that 
AS can be caused by disruption of the UBE3A gene and that other genes are likely to be 
involved in the UBE3A pathway. 34,35 The contribution of the other maternally expressed 
gene, ATP IOC to the AS phenotype is not well understood. 36 

The imprinting center mutations from several PWS patients were mapped to 
identify the smallest region of deletion overlap for PWS (PWS-SRO). 13 37 The PWS- 
SRO is located 5' of the SNRPN gene, including the first exon of the gene. The PWS- 
SRO is often referred to as the PWS imprinting center or PWS-IC. Although the PWS- 
SRO has been narrowed to approximately 4.3 kb, the actual deletions that have been 
shown to cause PWS are larger. 37,38 A smallest region of deletion overlap has also been 
identified for AS, and is referred to as the AS-SRO. The AS-SRO lies around 35 kb 
upstream of the PWS-SRO, and has been narrowed to 0.8 kb. 37,38 AS imprinting 
mutations never appear to overlap the PWS-SRO, while PWS imprinting mutations can 
overlap the AS-SRO. This also indicates a bipartite structure to the imprinting center, in 
which the PWS and AS components work together to control imprinting in the 15ql 1- 
q!3 region. 



The imprinting mutation class of patients for PWS and AS also suggests that the 
PWS-IC is required for the paternal pattern of gene expression across this approximately 
3 Mb region, while the AS-IC is required for the maternal pattern of gene expression. 
Human chromosome 15ql l-ql3 contains at least six paternally expressed genes or 
transcripts including MKRN3, MAGEL2, NDN, SNRPN, HBII-13, HBII-436, HBII-437, 
HBII-438A, HBII-438B, HBII-85, HBII-52, and the Lr££5A-antisense transcript 1718 21 - 39 - 44 
and two maternally expressed genes, UBE3A and ATP IOC. 19,36 

The PWS/AS region lies in a region of synteny with a region of central mouse 
chromosome 7, 17 and the gene order and structure is highly conserved. In addition, two 
mouse models— one a model for uniparental disomy, 21 and the other a targeted mutation 
in the PWS-IC 22 — suggest that the imprinted regulation of the genes in this region is also 
conserved. Mice are also capable of showing physical phenotypes such as obesity, 
failure to thrive, and small stature as well as behavioral phenotypes such as obsessive- 
compulsive disorder. Making the murine system a robust one for the study of the 
molecular and phenotypic aspects of PWS and AS. 

This dissertation focuses on the mechanisms and phenotypes of PWS. The rich 
genetic conservation between mouse and human has been harnessed to investigate these 
aspects of the disorder. This dissertation represents the four major accomplishments. 
First of all, the maternally expressed Ube3a gene was shown to be negatively regulated 
by an antisense transcript that is, in turn, positively regulated by the PWS-IC. This led to 
an attractive model for gene regulation in this region. Secondly, transgenes that show 
correct imprinted expression were identified, thus defining the minimal area sufficient to 
confer appropriate imprinted expression in vivo. These transgenes also provided 



8 

evidence for the functional conservation of the AS-IC in mouse as well as its approximate 
location in the region. Thirdly, the PWS phenotype in mice is influenced by different 
genetic backgrounds in mice. This discovery revealed a different phenotype in the PWS 
mouse model, as well as a method for identifying the particular modifying genes. 
Finally, transgenic rescue has identified a gene that is responsible for the small stature 
phenotype in the PWS mouse model and revealed a phenotype associated with snoRNAs. 
For ease of reading, each chapter contains sufficient background and methodology so that 
the reader may understand each one independently. 



CHAPTER 2 
IMPRINTED REGULATION OF UBE3A 

Introduction 

Angelman syndrome (AS) and Prader-Willi syndrome (PWS) are caused by 
defects in genes subject to genomic imprinting. AS results from a lack of a maternally 
inherited pattern of gene expression of human chromosome 15ql 1-ql 3, whereas PWS 
results from a lack of a paternally inherited pattern of gene expression in thel5ql 1-ql 3 
region. 45 The imprinted genes involved in AS and PWS are regulated by a bipartite 
imprinting center (IC) located upstream of the SNRPN gene. 26 The IC is divided into the 
Angelman syndrome imprinting center (AS-IC) and the Prader-Willi syndrome 
imprinting center (PWS-IC). Many of the mouse orthologs of these imprinted genes are 
located in the syntenic region of murine chromosome 7, l7 - 40 ' 42 46 - 48 furthermore, it is 
known that the PWS-IC is functionally conserved in the mouse. 22 

Patients with PWS lack expression of multiple paternally expressed genes. 
i5.i6.i8,4(M4 pws pat i ents can b e divided into 3 genetic classes: (1) maternal uniparental 
disomy for chromosome 15; (2) 3-4 Mb deletion of the paternally inherited 15ql l-ql3, or 
(3) imprinting mutations that are due to deletions of the PWS-IC and result in the 
silencing of the paternally expressed genes. 649 " 51 This latter class of patients indicates that 
the PWS-IC is required for the paternal pattern of gene expression across this 
approximately 2 Mb region. 



9 



10 

AS patients can be divided into five genetic classes: (1) paternal uniparental 
disomy for chromosome 15; (2) 3-4 Mb deletion of the paternally inherited 15ql l-ql3; 
(3) imprinting mutations that result from microdeletion of the AS-IC and subsequent 
disruption of the maternal pattern of gene expression; (4) biparental with a mutation in 
the maternally derived UBE3A gene; or (5) biparental with another as yet unidentified 
anomaly that cannot be described as an imprinting mutation or deletion. 6 34,50>52 It is the 
fourth, biparental class that has demonstrated that mutations in the UBE3A are sufficient 
to cause AS. 

Contrary to the known paternally expressed transcripts in the region, UBE3A does 
not exhibit strict imprinted expression: UBE3A is expressed predominantly from maternal 
chromosome 15 in the brain but shows biallelic expression in fibroblasts and 
lymphoblasts. 19,53 Albrecht et al. demonstrated that while the murine Ube3a gene is 
expressed exclusively from the maternal allele in hippocampal neurons and Purkinje 
cells, it is expressed from both alleles in other regions of the mouse brain. 48 Another 
difference between paternally expressed transcripts and UBE3A is that for most of the 
paternal genes, a region of differential methylation exists between the maternal and 
paternal alleles. I3,54-55 This is not the case for UBE3A, as to date, no differentially 
methylated CpG sites have been shown to exist near UBE3A (A. Lossie and D.J. Driscoll, 
personal communication). Together, these differences in imprinted expression and 
parental-specific methylation suggest that the mechanism that regulates the paternal 
program of gene expression in this region may not be the same mechanism as that which 
regulates UBE3A imprinted expression. 



11 

Rougeulle et al. reported the existence of an antisense transcript to the UBE3A 
gene. 41 Using RT-PCR of RNA derived from AS and PWS patient brains, they 
determined that this antisense transcript was imprinted, expressed only from the paternal 
allele. Rougeulle et al. suggested that the maternal-only expression of UBE3A may result 
from tissue specific expression of this paternally expressed antisense transcript. 
Therefore, using a PWS-IC deletion (APWS-IC) mouse model, 22,55 the role of the PWS-IC 
in the regulation of the murine Ube3a antisense transcript and its subsequent effects on 
the parent-of-origin specific expression of the sense Ube3a transcript were investigated. 
It was found that the murine Ube3a transcript is expressed exclusively from the paternal 
allele, the antisense transcript is regulated by the PWS-IC, and the level of paternal vs. 
maternal Ube3a expression is different between APWS-IC mice and wild-type 
littermates, suggesting a role for the antisense transcript in the regulation of maternal 
Ube3 expression 

Materials and Methods 

Strains and Matings 

Mouse strains used were mus musculus 129/Sv (129/Sv or D) and a strain of 
C57BL/6J that is congenic for the PWS/AS syntenic region of chromosome 7 from mus 
castaneus (B6.CAST.c7 or C) 56 . For (B6.CAST.c7 X 129/Sv) F, matings, a female 
B6.CAST.c7 mouse was mated with a male APWS-IC mouse on a 129/Sv strain 
background. Both APWS-IC mice and wildtype littermates were used. The (129/Sv X 
B6.CAST.c7) F, mice were generated by mating a female 129/Sv mouse with a male 
B6.CASTc7 mouse. 



12 

Identification of Polymorphisms 

Total brain RNA was prepared from both 129/Sv mice and B6.CAST.c7 mice. 
Each RNA sample was used to create single-stranded cDNA using random primers. 
These cDNAs were used to program PCR reactions using primers corresponding to 
Ube3a exons 4 and 7 (5 ' -CCTGC AG ACTTGAAG AAGC AG-3 ' and 5'- 
GAAAACCTCTGCGAAATGCCTT-3 '). The resulting products were cloned, sequenced 
and compared. A polymorphism found in exon 5 created a Tsp5Q9I restriction site in the 
129/Sv clone and a Bst4Cl site at the same location in the B6.CAST.c7 clone. 
RT-PCR 

Total RNA was isolated from total brain obtained from neonatal mice. RNA was 
extracted using RNAzol (Tel-Test, Inc.) according to instructions. This total brain RNA 
(10u.g) was then pretreated with DNAse I (Life Technologies) and half of the reaction 
was subsequently used to synthesize first strand cDNA with Superscript II reverse 
transcriptase (RT) (Life Tehcnologies) and either strand-specific primer 5F or random 
primers (Life Technologies). The other half of the reaction was manipulated in parallel in 
the absence of RT. One-twentieth of the +RT or -RT reactions were used to seed PCR 
reactions using the following conditions: lOmM Tris-HCl, pH 8.3, 50mM KC1, 0.125mM 
of the four dNTPs, 1 unit Taq DNA polymerase (Boehringer), and 4u.M of each of the 
appropriate primers. Sequences of the primers were 5F (5'- 

C AC ATATGATGA AGCTACG A-3 ' ), 5iR (5 ' -C AG AA AG AG AAGTG AGGTTG- 3 ' ) , 
and 6R (5'-CACACTCCCTTCATATTCC-3'). PCR amplification conditions were: 95°C 
for 5 min, followed by 30 cycles of 94°C for 30 s, 60°C for 45 s, and 72°C for 45 s. The 
final cycle was followed by an extension step for 10 min at 72°C. The PCR products 



were either left undigested or digested with enzymes and analyzed on either a 2% or 
4.8% agarose gel. 

Results 

The expression of a Ube3a antisense transcript was first verified in mouse brain 
by performing PCR across intron 5 using randomly primed murine brain cDNA. Two 
bands were amplified from the cDNA: the smaller correctly spliced Ube3a transcript and 
the larger unspliced form of the Ube3a gene, which was verified by sequencing (data not 
shown). The existence of this unspliced form is consistent with a Ube3a antisense 
transcript that is expressed in mouse brain. The presence of the antisense transcript was 
then confirmed, and its paternal specific expression was determined using F ( progeny 
derived from reciprocal matings between wild-type 129/Sv mice and B6.CAST.c7 mice 
which are congenic for the PWS/AS syntenic region from mus musculus castaneus 
chromosome 7 on a C57BL/6J background. 56 DNAsel treated RNA prepared from 
brains of F, mice was used to program strand-specific reverse transcription reactions 
using a forward primer from exon 5 (5F; Fig. 2-1 A), thus yielding cDNA made from the 
antisense transcript. Using this strand specific cDNA as a template, PCR was performed 
with the forward primer 5F and a reverse primer from intron 5 (5iR; Fig. 2-1 A). A 
polymorphism in Ube3a exon 5 between Mus musculus domesticus and Mus musculus 
castaneus enabled the determination of the parent of origin by digestion of the RT-PCR 
products with the enzyme 7s/>5091, which digests only the domesticus allele. The RT- 
PCR product derived from the antisense transcript was found to be resistant to 7s/?509I 
digestion when the father was B6.CAST.c7, but was cleaved by this enzyme when the 



father was 129/Sv. This demonstrates that the antisense Ube3a transcript is expressed 
exclusively from the paternally inherited allele (Fig. 2- IB). 

As previous studies have shown that all other paternally expressed transcripts in 
the region are regulated by the PWS-IC in humans and mice, 6 - 12 - 13 - 20 - 22 - 55 it was necessary 
to determine whether the paternally expressed Ube3a antisense transcript was also 
subject to regulation by the PWS-IC. Therefore, the PWS-IC deletion (APWS-IC) mouse 
strain was used. This mouse strain was created by targeted deletion in ES cells that 
includes exons 1 through 6 of Snrpn and extends approximately 23 kb 5' of the gene. 22,55 
B6.CAST.c7 females were mated with APWS-IC males and total brain RNA was isolated 
from the resulting newborn (B6.CAST.c7 X APWS-IC) F, mice. The RNA was DNAsel 
treated and used to direct strand- specific RT reactions, each containing the sense 5F 
primer and an antisense fi-actin primer. The resulting cDNA was used as a template for 
PCR using the primers 5F and 5iR, or two fi-actin primers. Figure 2-2 shows that the 
Ube3a antisense-derived 840 bp band present in the wild type cDNA is absent in the 
APWS-IC littermate cDNA. In contrast, actin is detected in both cDNA preparations. 
This result demonstrates that the Ube3a antisense transcript is positively regulated by the 
PWS-IC. 

Finally, to determine if the paternal Ube3a antisense transcript regulates the level 
of paternal Ube3a mRNA, RT-PCR followed by restriction enzyme digestion was 
performed to examine the levels of Ube3a expression produced from each parental allele 
in both wild-type and APWS-IC mice. To distinguish between Ube3a expression derived 
from the two parental alleles, two exon 5 polymorphisms between Mus musculus 
domesticus (D) and Mus musculus castaneus (C) were used: the enzyme Bst4Cl cuts the 



15 

castaneus allele once but does not cut the domesticus allele; and the enzyme 7s/?509I cuts 
the domesticus allele twice and the castaneus allele only once (Fig. 2-3A). The 
experimental cross of B6.CAST.c7 X APWS-IC was established to obtain both wild-type 
F ( pups (C X D wt) and APWS-IC F, pups (C X D A). The control strain reciprocal cross 
of 129/Sv X B6.CAST.c7 was also established to obtain wild-type F, pups (D X C wt). 
After birth, the pups were sacrificed, total brain RNA from all three classes was isolated, 
and then random primers were used to prepare cDNA. Subsequent PCR amplification 
with the Ube3a primers 5F and 6R demonstrated that Ube3a was detectable in all three 
classes of F, mice (Fig. 2-3B, first three lanes). Upon digestion of these RT-PCR 
products Bst4Cl, expression of both parental Ube 3a alleles was detected in all three 
classes. These results are expected, since Ube3a is known be expressed from both alleles 
in most regions of the brain, exhibiting imprinted expression in only a subset of brain 
cells. 48 However, in the PWS-AIC F, lane (C X D A) the ratio of paternal (uncut) to 
maternal (cut) Ube3a expression was found to be elevated compared to that observed in 
the wildtype F, lane (C X D wt). These results were confirmed by digestion of the RT- 
PCR products with Tsp509l. In this case, the paternal Ube3a allele (138 bp + 87 bp 
bands) is clearly present at significantly higher levels in the APWS-IC F, lane (C X D A) 
than the wildtype F, lane (C X D wt). These results demonstrate that the loss of paternal 
Ube3a antisense expression in APWS-IC mice is accompanied by an increase in paternal 
sense Ube3a expression. 

Discussion 

In this study, the murine Ube3a antisense transcript was shown to be exclusively 
expressed from the paternal allele and positively regulated by the PWS-IC. Furthermore, 



16 



mice that inherit a paternal deletion of the PWS-IC, and thus lack Ube3a antisense 
expression have an upregulated paternal Ube3a allele relative to the maternal allele, as 
compared to the wild-type controls. Together, these results strongly suggest that the 
paternally derived antisense transcript negatively regulates the level of paternally 
expressed Ube3a mRNA. 

The differences between the imprinted expression of paternally expressed genes 
in the region and UBE3A strongly suggest that the maternal-only expression of the 
UBE3A gene is regulated by a different mechanism. The hypothesis that the tissue 
specific imprinting of UBE3A is an indirect consequence of the imprinted antisense 
transcript has been previously proposed by this lab and others. 41,57 Specifically, in tissues 
that express the antisense transcript from the paternal allele, production of UBE3A 
mRNA from the paternal allele is inhibited (presumably this inhibition occurs at the 
transcriptional level). However, the maternal UBE3A allele is not prevented from 
producing mRNA, so in these tissues, maternal specific expression is observed (Fig. 2- 
4A). In tissues that do not express the antisense transcript but do express the UBE3A 
gene, biallelic expression of UBE3A is observed (Fig. 2-4A). The data presented here 
using the mouse model are consistent with this hypothesis and further suggest that the 
imprinted expression of UBE3A is not a direct consequence of the UBE3A promoter or 
the AS-IC. Rather, it appears to be the PWS-IC which negatively regulates the expression 
of paternal UBE3A via the antisense transcript. 

These data support a model in which the primary targets of imprinting in the 
PWS/AS region are the paternally expressed transcripts, one of which, the UBE3A 
antisense transcript, results in maternal-specific expression of the UBE3A gene. The 



advantage of this model is that it greatly simplifies the seemingly complex pattern of 
gene regulation in the PWS/AS region. A positive regulatory element, the PWS-IC, is 
responsible for establishing and maintaining the paternal mode of gene expression in 
somatic tissues (Fig. 2-4B). A negative regulatory element, which is assumed to be the 
AS-IC, serves to prevent activation of the paternal program on the maternally inherited 
chromosome, presumably by blocking PWS-IC function in the maternal germline (Fig. 2- 
4C). As a result, different patterns of gene expression are observed for the two parental 
alleles. The paternally expressed genes including the UBE3A antisense transcript are 
expressed from the paternal chromosome but not from the maternal chromosome. In 
contrast, UBE3A is "constitutively" expressed from the maternal chromosome in all cell 
types that exhibit UBE3A expression, whereas the paternal UBE3A allele is only 
"allowed" to be expressed in cell types that do not express the antisense transcript. 
Therefore, according to this model, it is only the paternally expressed genes that are 
subject to regulation by the IC as a whole. This reduces the complex regulation in this 
region to a simple ON/OFF decision: the PWS-IC operates in the ON mode unless shut 
OFF by the AS-IC element. Therefore, the key to understanding imprinting in this region 
will be to determine how the AS-IC inhibits the PWS-IC on the maternal chromosome. 

Finally, as the antisense transcript appears to be the cause of the tissue specific 
imprinting of UBE3A, it will be interesting to determine what regulatory elements result 
in the expression of the antisense in such a restricted cell population in the brain. It is 
possible that there are positive regulatory elements that drive expression of the antisense 
transcript in only a few cell types. Alternatively, there may be negative regulatory 
elements that prevent expression of the antisense in most cell types. At this point, it can 



18 

only be said that the PWS-IC exerts a positive effect on the transcription of the antisense. 
There may be additional levels of control that restrict the antisense to fewer cell types 
than the other PWS-IC regulated paternally expressed genes in the region. 



5F 



7jp509I 



C (B6.Cast.c7) 



21 bp^ 



6R 



640 bp 



179 bp 



D(129/Sv) 21 bp a 87 bp 



f »/ bp -f- 



553 bp 



179 bp 



B 



uncut 



i r 



7ip509I No RT 
1 I 1 



maternal 
paternal 



C 
X 
D 



D 

X 
C 



C D 



C 
X 
D 



D 

X 
C 



C D 



C 
X 
D 



D 

X 

c 



840 bp 




19 





Exon 5 




Exon 6 






<- 






5iR 


<- 



Figure 2-1. The antisense transcript is expressed from the paternally-derived 
allele only. A. Map of the exon 5 to exon 6 region of Ube3a and schematic 
showing expected PCR products, restriction sites, and fragment sizes. Triangle 
lollipops represent 7^5091 sites, and horizontal arrows represent primers used 
for RT and PCR. B. The antisense transcript was amplified by RT-PCR from 
neonatal brains from (B6.CAST.c7 X 129/Sv) F, and (129/Sv X B6.CAST.c7) Fj 
mice, and by PCR from genomic DNA from parental B6.CAST.c7, and 129/Sv 
mice using primers 5F and 5iR. The products were digested with Tsp509I and 
electrophoresed along with the uncut controls on a 2% agarose gel. C indicates 
the B6.CAST.c7 strain, while D indicates the domesticus strain, 129/Sv. Arrows 
indicate pertinent fragments and their sizes. 




Figure 2-2. The antisense transcript is not expressed in the APWS-IC 
mouse. cDNAs were made from a APWS-IC mouse and a wildtype 
littermate using primers 5F and the 3' fi-actin control. PCR was done 
using primers 5F and 5iR. The resulting products were run on a 2% 
agarose gel. 



C (B6.Cast.c7) 



5F 

»t T 



Exon 5 



Exon 6 



6R 



246 bp 



108 bp 138 bp 



21 bp 225 bp 



uncut 

Bst 4CI 
7sp509I 



5F 



D(129/Sv) 
_f_ 



Exon 5 



Exon 6 



6R 



246 bp 



246 bp 
t t 



21 bp 87 bp 138 bp 



B 



maternal 

paternal 
genotype 



Bst4Cl 



7>p509I 



No RT 



246 bp 

138 bp 
108 bp 




Figure 2-3. Paternal Ube3a expression is increased in APWS-IC mice. A. 

B6.CAST.c7 and 129/SvEv alleles are diagrammed showing PCR products, 
restriction sites, and fragment sizes. The circle lollipop indicates the location of 
a Bst4Cl site, while the triangle popsicle indicates the Tsp509l sites. Horizontal 
arrows show primers used for RT and PCR. B. Sense Ube3a cDNA was made 
using RNA derived from (B6.CAST.c7 X 129/Sv) F,, (B6.CAST.c7 X APWS- 
IC) F„ and (129/Sv X B6.CAST.c7) Fj brains using random primers and PCR 
amplified using primers 5F and 6R. The spliced, sense band was gel purified 
and each sample was divided three ways: one portion was left uncut, the second 
was cut with BstACl, and the third was cut with Tsp5091. These samples were 
loaded on a 4.8% agarose gel. C indicates the B6.CAST.c7 strain, while D 
indicates the domesticus strain, 129/SvEv. For the genotype, wt indicates a wild- 
type mouse; A indicates a APWS-IC mouse. Arrows indicate the location of the 
fragments and their sizes. 



22 



Mkrn3,Magel2, Ndn 



Snrpn 



Ube3a AtplOC 






Figure 2-4. Model of gene regulation in the PWS/AS region. A. In the 

brain, two different patterns of UBE3A expression occur. In cell types that 
express the paternal antisense transcript, UBE3A is imprinted (top box). In 
cell types that do not express the antisense transcript, UBE3A is biallelically 
expressed (bottom box). B. On the paternally-inherited chromosome, the 
PWS-IC establishes the paternal mode of gene expression. C. On the 
maternally-inherited chromosome, the AS-IC prevents activation of the 
paternal program, presumably by blocking PWS-IC function. 



CHAPTER 3 
IMPRINTED TRANSGENES 

Introduction 

Use of Transgenic Mice to Study Gene Regulation 

Transgenes have been used to study regulation of gene expression in many 
systems. They serve to isolate regulatory elements and relocate them to an ectopic 
location in the genome where their local activities may be separated from the influence of 
adjacent cooperating regulatory elements. One of the simplest instances of using 
transgenes to study regulatory elements involved the tyrosinase gene. 58,59 A 250 kb yeast 
artificial chromosome (YAC) was found to faithfully express the tyrosinase gene. 58 This 
YAC was subsequently engineered to harbor a deletion of a DNasel hypersensitive site, 
demonstrating that the site is necessary for correct expression of the tyrosinase gene. 59 
The regulation of the (3-globin locus is another example that has benefited from the use of 
transgenes. In the (3-globin locus, long distance regulatory elements determine the tissue 
specificity and developmental timing of the expression of various exons in the locus. 60 
Similar to the tyrosinase locus, DNasel hypersensitive sites were deleted from a YAC 
using homologous recombination, and were determined to be locus control regions in 
vivo by demonstrating that YACs that were missing or had duplications of these 
hypersensitive sites no longer showed correct temporal expression of the (3-globin exons. 
This was accomplished by making transgenic mice that contained single copies of the 
wild-type or modified YAC. 61 ' 62 



23 



Imprinted regions are also good candidates for study using large transgenes. 
Wutz and colleagues used YAC transgenes to examine the role of chromosome location 
and the role of an intronic CpG island in the imprinted expression of Igf2r, a maternally 
expressed gene. 63 They were able to show that lgflr imprinting is not dependent on 
chromosome location, but that it is completely dependent on the intronic CpG island, 
such that deletion of the CpG island causes derepression of the paternal Igflr allele. 
Regulation of another well understood imprinted region, the H19-Igf2 locus, has also 
been studied using transgenic mice. The Tilghman lab were able to show that multiple- 
copy insertions of an H19 transgene could exhibit parent-of-origin dependent expression 
patterns as well as proper DNA methylation patterns, but these patterns were not 
maintained when the same transgene was present in single-copy. 64,65 The authors 
reasoned that the transgene constructs lacked sufficient sequence information to confer 
proper imprinted expression. Kaffer et al. subsequently identified a 137 kb bacterial 
artificial chromosome (BAC) that exhibited the proper pattern of imprinted gene 
expression for the H19-Igf2 locus in single-copy transgenic lines. Hark et al. also used 
the smaller transgenes that were not imprinted in single copy to test a model of gene 
regulation for the Igf2 gene. 66 The authors used a reporter construct that contained a 
portion of the imprinted control region (ICR) followed by a polymorphic copy of the H19 
gene and its enhancers derived from Mus musculus spretus, another intact ICR, and a 
luciferase gene. In subsequent transgenic lines a portion of the ICR that acted as a 
putative barrier element, separating H19 from its enhancers, was deleted. These lines 
showed higher expression of luciferase, indicating that the barrier element acted as 
hypothesized. 66-69 



25 

Single copy transgenes could be used to elucidate some of the elements necessary 
to achieve proper imprinted expression in the PWS region. Identification of a transgene 
that is capable of demonstrating correct imprinted expression in single copy would 
elucidate the minimal genetic information that is sufficient for imprinting in this region. 
This transgene could then be used to make various permutations and delineate whether 
individual components are necessary for either paternal expression or maternal 
repression. Furthermore, if an efficient method for modifying the transgene were 
developed, a latent mark could be applied to the transgene that would make it easy to 
distinguish between transgenic and endogenous expression. The system could then be 
used to easily introduce new changes into the transgene. 

Previously, our lab produced seven lines of transgenic mice carrying either a 
human or a mouse PI clone. The human PI clone was chosen because it was known to 
include the human Angelman syndrome smallest region of deletion overlap (AS-SRO) as 
well as the Prader-Willi syndrome smallest region of deletion overlap (PWS-SRO). 
Furthermore, it was polymorphic to the homologous mouse locus that is syntenic to 
15ql l-ql3. 70 The expressed polymorphisms between the human PI clone and the mouse 
locus would allow us to easily distinguish expression originating from the transgene from 
that coming from the endogenous locus. This human PI clone was never correctly 
imprinted in 5 transgenic lines produced from it. The human PI clone was expressed 
when it was inherited from both father and mother, suggesting that while the human 
PWS-IC positive element is functional in mouse, the human AS-IC element, which is 
predicted to be a negative regulatory element, does not behave as one in mouse. The 
mouse PI clone, on the other hand, contained the Snrpn gene as well as the functionally 



26 

defined PWS-IC, but since the AS-IC has not been functionally or physically identified in 
mouse, we could not be certain whether or not it contained the AS-IC. Two lines were 
created from the murine PI clone. One was multi-copy and showed correct imprinted 
expression, while the other was single copy and was expressed regardless of the parent- 
of-origin. It was hypothesized that the mouse PI clone did not contain the entire AS-IC, 
indicating that more genetic information, specifically 5' sequence, is necessary to imprint 
a transgene in single copy. 

Two murine bacterial artificial chromosome (BAC) clones were identified that 
were capable of being expressed in the appropriate parent-of-origin dependent manner. 
These BAC clones specify a minimal region that is sufficient to confer correct imprinted 
expression at an ectopic location in the genome. 

Materials and Methods 

Screening the BAC Libraries 

The Research Genetics (Huntsville, AL) BAC library was initially screened for 
BAC clones. This library was made from the 129/Sv strain of mouse. The library 
consisted of 9 filters, each carrying 27,648 unique clones with an average insert size of 
130 kb spotted in duplicate onto 9 different membranes. A PCR-generated probe that 
included the first exon of Snrpn and a 2.2 kb EcoRl-EcoRW restriction fragment that was 
5 kb 3' of Snrpn was used to screen the library. 

Prior to hybridization, BAC membranes were washed with 1500 mL of 6X SSC 
and 0.1% SDS at room temperature for 15 minutes. The membranes were then rinsed 
twice with 1500 ml of 6X SSC for 15 minutes at room temperature, and hybridized 
according to the Research Genetics protocol. Briefly, the membranes were prehybridized 



27 

in roller bottles with 3 membranes per bottle, each separated by one sheet of Flow Mesh 
(Diversfied Biotech, Boston, MA), in 120 ml of HyperHyb (Research Genetics, 
Huntsville, AL) per bottle at 65° C for 20 minutes. At least 10 7 counts of each probe 
were boiled for 5 min, snap cooled on ice, and then added to 3 ml of HyperHyb that was 
pre- warmed to 65° C. One ml of hybridization solution was added to each of the roller 
bottles, and the membranes were allowed to hybridize for 2 hours at 65° C. 

The membranes were washed in the roller bottles three times for 15 minutes at 
65° C with 30 ml IX SSC and 0.1% SDS. They were then removed from the bottles and 
washed twice more at 65° C with 1000ml of IX SSC and 0.1% SDS for 15 minutes. 
Finally, the membranes were rinsed with IX SSC at room temperature. The membranes 
were then wrapped in cellophane and exposed to film (XAR, Kodak) overnight at -80° C. 
Positive clones were identified following the manufacturer's instructions, and the 
following clones were ordered: 181J3, 573B8, 518H22, and 397F16. 

The Roswell Park Cancer Institute (RPCI) RPCI-23 BAC library was also 
screened for potential imprinted transgenes using two probes; the 2.2 kb Eco RI- EcoRW 
fragment lying 5 kb 3' of Snrpn and a 1 kb Sacl-Notl fragment that lies 30 kb 5' of 
Snrpn. This library is made from the C57BL/6J strain of mouse, with an average insert 
size of 200kb and is the template library for the NIH mouse genome sequencing efforts. 
The RPCI-23 library consists of 10 filters, each representing 18,432 independent clones 
that have been spotted in duplicate. To screen this library, the filters were first rinsed for 
10 minutes at room temperature in 6X SSC. The BAC library filters were placed into 
roller bottles, five membranes per bottle, rolled with intermittently with sheets of Flow 
Mesh. The filters were pre-hybridized for 2 hours at 65° C with 200 ml of Church buffer 



28 

(1 mM EDTA; 0.5M NaHP0 4 , pH 7.2; 7% SDS; and 1% BSA). Approximately half of 
the pre-hybridization solution was poured off. The boiled, snap cooled probes (at least 
10 7 counts, each) were then added, and the membranes were hybridized overnight at 
65°C. 

The next morning, the membranes were first washed 5 times for 10 minutes each 
in the bottle at 65° C with 200 ml per tube of Wash I (1 mM EDTA; 40 mM NaHP0 4 , pH 
7.2; 5% SDS; and 0.5% BSA). The filters were then removed from the bottles and 
washed 3 times for 15 minutes each at 65° C with Wash II (1 mM EDTA; 40 mM 
NaHP0 4 , pH 7.2; and 1% SDS). The membranes were wrapped in cellophane and 
exposed to film overnight and for 3 hours each at -80° C. Positive clones were identified 
following the manufacturer's instruction and ordered from RPCI. The following BAC 
clones were obtained from this screen: 73N19, 78014, 148C8, 160G24, 227M21, 
284G21, 359M8, 365C3, 382D19, 438B21, 396A12, 421K3, 9715, 215A9, 264P24, 
276M11.425D18, 380J10, 391B10, and 437G6. 
Probe Preparation 

Probes were labeled using the Prime-It II kit (Stratagene, La Jolla, CA), with 
minor modifications. Probe DNA (200 ng) was boiled for 5 minutes in the presence of 
random hexamers. The DNA-hexamer mix was cooled on ice, and 5X reaction buffer 
was added along with 25 uGi of aP 32 dCTP and 1U of exo" klenow. The labeling reaction 
was placed at 37° C for 15 minutes. After labeling was complete, the probe was purified 
using the nucleotide removal kit (QIAGEN) and eluted in 200 ul of water. Purified 
probes were boiled for 5 minutes and snap cooled on ice prior to use in hybridization. 



29 

Verification of BAC Clones 

Single BAC colonies were tested to verify that the clone represented a BAC insert 
derived from the Snrpn locus. Four single colonies from each BAC were streaked in 
duplicate onto Luria broth plates that were supplemented with chloramphenicol (LB-CA). 
The plates were overlayed with a nylon filter and grown overnight at 37° C. The next 
morning, the bacteria were lysed on the filters as follows: first, the filters were soaked in 
alkali solution (1.5 M NaCl and 0.5 N NaOH) for 4 minutes, followed by neutralizing 
solution (1.5M NaCl and 1M Tris, pH 7.4) for 4 minutes, and finally rinsed in 6X SSC 
for 1 minute. The filters were then pre-hybridized in Church buffer for one hour at 65° 
C, and then one plate was hybridized with the 5' Snrpn probe, while the other was 
hybridized with the 3' Snrpn probe. Upon washing and laying the filters to film, the 
results were used to determine whether or not the clone was derived from the Snrpn locus 
and whether or not the clone encompassed the 5' or 3' flanking regions of Snrpn. 
End Sequencing BAC Clones 

The BAC clones were end sequenced using the big dye termination reaction to 
obtain sequences unique to the mouse genomic DNA insert. Reactions were performed 
on 1 mg of BAC DNA that had been sheared by passing it five times through a 25 gauge 
needle using a modified version of the manufacturer's instructions, with 12ul of 
sequencing buffer, 4 u.1 of big dye, 16 pmol of primer, and 2.5mM MgCl 2 . The reactions 
were subjected to an initial 5 minute cycle at 95° C, followed by 30 cycles of 95° C for 
30 seconds, 50° C for 10 seconds, and 60° C for 4 minutes. The completed reactions were 
purified from free nucleotides using Performa DTR Gel Filtration Columns (Edge 



30 

Biosystems), dried, and sequenced by the University of Florida Center for Mammalian 
Genetics Sequencing Core. 
Pulse-Field Gel Electrophoresis 

Pulse-field gel electorphoresis was carried out to create a rough restriction map 
and tiling pathway for the BAC clones. BAC DNAs digested with the appropriate 
enzymes were electrophoresed on a 1% genetic technology grade agarose gel (Nusieve, 
FMC) made with IX TBE buffer. Samples were mixed with 10X ficoll loading dye and 
ran at 200V with a switch time of 1-12 seconds for 12.5 hours. This provided resolution 
of bands from 6-200 kb. The gel was stained with a dilute ethidium bromide solution, 
photographed, and blotted as described below. 
Southern Blot 

Agarose gels carrying 5 u.g of digested genomic DNA or 1 \ig of digested BAC 
DNA were photographed and UV nicked for 5 minutes. The gel was then soaked in 
alkali solution for 45 minutes, followed by soaking in neutralizing solution for 90 
minutes. The gel was then transferred to Hybond nylon membrane (Amersham) in 10X 
SSC overnight. The membrane was then rinsed in 2X SSC and baked at 80° C for 2 
hours. Hybridization was carried out according to the method of Church and Gilbert. 
Membranes were prehybridized in 20 ml of Church buffer for 1 hour at 65° C. The 
prehybridization solution was poured off and replaced with 5 ml of fresh solution and the 
boiled and cooled probe. Hybridization continued overnight at 65° C. Washing was 
carried out three times in 2XSSC and 0.1% SDS at 65° C for 15 minutes each. 



31 

Transgenic Mice Production 

For production of transgenic mice, BAC DNA was prepared by large scale 
plasmid preparation, using the alkaline lysis method. First, log phase cultures of BAC 
culture grown in LB-CA were centrifuged for 5 minutes at 6,000 rpm to pellet the cells. 
The supernatant was poured off and the pellet was resuspended in GTE solution (50 mM 
glucose; 25mM Tris, pH 8.0; and 10 mM EDTA) plus lysozyme. The suspension was 
then lysed in 0.2 N NaOH and 1% SDS. After addition of the lysis solution, the lysate 
was neutralized in 0.5 M potassium acetate, and poured through cotton gauze into Oak 
Ridge tubes. BAC DNA was precipitated with ethanol, and the pellet was resuspended in 
4 ml TE with 10 u.g/ml RNAse A, and allowed to incubate at 37° C for 1 hour. Cesium 
chloride and ethidium bromide were added to the DNA solution, and the solution was 
loaded into a heat sealable tube. The tube was sealed and centrifuged at 58,000 rpm for 
18 hours. BAC supercoils were viewed with a UV light, and both nicked and supercoiled 
bands were pulled. The DNA was extracted with tert-butanol until the ethidium bromide 
was completely depleted from the BAC DNA. The DNA was then dialyzed twice against 
1 L TE buffer, using dialysis tubing for 6 hours each. The concentration of the BAC 
DNA was measured by spectrophotometer, and then diluted to 2.5 ng/u.1 for 
microinjection. 

Transgenic mice were made by pronuclear injection into oocytes that were 
obtained from superovulated, fertilized, FVB/NJ female mice. The fertilized, injected 
oocytes were then implanted into (B6D2) F, females that were made pseudopregnant by 
mating them with vasectomized (B6D2) F, males. 



Mouse Husbandry 

At 3 weeks of age approximately 2mm of tail was clipped from each mouse, and 
the ear was punched for identification. Genomic DNA was prepared from the tail piece 
by incubation in tail lysis buffer (100 mM Tris, pH 8.5; 5 mM EDTA; 2% SDS; and 
200mM NaCl) with 100 ng/ul proteinase K overnight at 55° C. The tail lysate was 
extracted with phenol: chloroform: isoamyl alcohol (25:24:1) and then precipitated with 
ethanol. The tail DNA was screened for the presence of the transgene by PCR using both 
end sequences from the BAC. 
Mouse Strains 

APWS-IC mice maintained on the C57BL/6J background were developed by 
Yang et al. and constituted a 35 kb deletion encompassing 16 kb of sequence 5' of Snrpn 
and 19 kb 3' of Snrpn 22>55 . Lines 215A, 215B, 215C, 425A, and 425B were all transgenic 
lines produced as described above and maintained on the FVB/NJ strain background. 
Transgenic line 1707 was previously described and was originally maintained on the 
DB A/2J strain background 10 . For the purposes of this study, this line was backcrossed 
for 10 generations onto the FVB/NJ strain background. 

Results 

YAC Studies 

A YAC that contained 320 kb of genetic information spanning the Snrpn area (Fig 
3-1) was previously identified. Pulse-field gel electrophoresis was used in conjunction 
with rare cutting enzyme digests to confirm that the YAC was indeed intact and 
contained the genetic information that we anticipated. Several unsuccessful attempts 
were then made to subclone the regions that comprised the junction between the YAC 



vector arms and the mouse DNA insert. Despite the failure to find these junction 
fragments, an attempt was made to make transgenic mice using the YAC, but it was not 
successful. Ultimately, the attention was focused on BACs, since candidate clones for 
imprinted transgenes had been identified with our screen. It was surmised that the YAC 
was not ideal for use, since a YAC of that size is difficult to characterize and manipulate. 
129/Sv BAC Library Screen 

B AC clones were the best suited for transgenic studies, since they are usually 
larger than the PI clones that were previously shown not to contain sufficient sequence 
information to be imprinted in single copy. While they are larger than PI clones, BAC 
clones are smaller than YACs and are maintained as supercoils, making them more 
resistant to shearing. They also have the advantage over YACs of being propagated in E. 
coli rather than S. cerevisiae, which makes their isolation and handling much easier. Most 
importantly, BACs are propagated in RecA " cells, and are less likely to suffer 
rearrangement during manipulation and propagation than YACs. Traditionally, the 
advantage of using YACs as opposed to BACs was the ability to make modifications by 
exploiting the efficient homologous recombination in S. cerevisiae. However, early in 
this project a preliminary method of BAC modification using homologous recombination 
was published, making it possible to make similar changes to BAC clones 71 . 

The first BAC library that we screened was derived from the 129/SvEv strain of 
mouse, and had an average insert size of 130 kb. Using a mixture of the Snrpn exon 1 
probe and a 2.2 kb EcoRl-EcoRV that lies 5 kb 3' of Snrpn as probes in this screen, four 
BACs were isolated: 181J3, 573B8, 518H22, and 397F16. These BAC clones were each 
streaked to LB-CA plates to obtain single colonies. Four single colonies representing 



34 

each BAC clone were grown in 3 ml overnight cultures and DNA was isolated from each 
culture. The BAC DNA was cut with EcoRl, and a Southern blot was performed. The 
blot was first probed with a 1 kb Sacl-Notl fragment that yielded a large fragment about 
20 kb 5' of Snrpn, and later stripped and re-probed with the 3' Snrpn probe. Two BACs, 
397F16 and 518H22 spanned 5' Snrpn, but terminated short of the 3' Snrpn probe, while 
the other two BACs, 573B8 and 181J3, conversely hybridized with the 3' probe, but 
appeared to lack 5' sequences (Fig. 3-1). All four BACs were end sequenced. Next, the 
BAC clones were each cut with Sail, Notl, Sacll, or a combination of the enzymes and 
subjected to pulse-field gel electorphoresis and Southern blot. Using the 5' and 3' 
probes, probes intergenic to the Snrpn gene, and probes generated using the end 
sequences from each BAC, the approximate boundaries as well as a rough restriction map 
for each BAC clone was determined. Following this analysis, only two BACs, 397F16 
and 573B8 remained interesting as potential imprinted transgene candidates. The 
397F16 clone contained the most 5' sequence of the BAC clones, but it terminated before 
the polyadenylation signal of the Snrpn gene. This was a problem because it could cause 
an expressed Snrpn transcript to be targeted for degradation. The other BAC of interest, 
573B8 contained the entire Snrpn gene, but did not contain enough 5' sequence. Since 
neither BAC was ideal for use as an imprinted transgene, a plan was made to join these 
two BACs using the recently developed BAC modification scheme to produce the 
additional 5' sequence as well as the complete Snrpn transcript. 
RecA Mediated Homologous Recombination 

The first method attempted was analogous to the yeast two-step homologous 
recombination procedure, as described by the Nat Heintz group. 71 It revolved around 



35 

supplying the bacterial RecA gene, which was missing in the BAC host strain, on a 
selectable episome. 71 The bacterial RecA gene was provided on a plasmid with a 
temperature sensitive origin of replication. The plasmid could be positively selected for 
using tetracycline resistance. The plasmid could also be subjected to negative selection 
by plating on plates containing fusaric acid, since fusaric acid is toxic to cells that 
produce the tetracycline resistance gene. The idea behind this recombination scheme was 
that one could subclone the desired modification or sequence change flanked by 5' and 3' 
homology arms of at least 500 bp, into this temperature sensitive recombination vector. 
When transformed into the BAC strain, positive selection for the modification construct 
would identify colonies in which insertion of the construct had occurred. Then, negative 
selection with fusaric acid would cause resolution of the integrant into one of two 
possible conformations. One possible conformation is that in which the desired change 
has been made, and the other conformation recovers the unmodified BAC. 

The scheme to join the two pertinent BACs involved using homologous 
recombination to insert a neomycin resistance cassette flanked by loxP sites 
approximately 30 kb upstream of the Snrpn gene on the 397F16 BAC. A 35 kb Sail 
fragment from the modified 397F16 BAC could then be excised and recombined into the 
573B8 BAC using one arm of homology anchored in the BAC vector and the other end 
anchored near the first exon of Snrpn. The neomycin resistance gene could then be 
deleted using ere recombinase, leaving a single loxP site. The resulting BAC would have 
35 kb sequence 5' of Snrpn and 55 kb 3' of Snrpn, and would be a good candidate for an 
imprinted transgene (Fig 3-2). While this method has been successful in other labs, it 
was not so in this case due to the difficulty of subcloning the initial recombination 



construct into a low-copy number plasmid that is temperature-sensitive and already quite 
large (llkb). 

RecET Mediated Homologous Recombination 

The second method, pioneered by Francis Stewart's group, was based on the 
RecET system in lambda phage. 72 This system is reputedly much more efficient than 
that of RecA and requires smaller homology arms. The RecET mediated recombination 
can be carried out using only large oligos and the RecET plasmid electroporated into the 
BAC containing bacterial strain. The RecE and RecT gene are expressed from the pBAD 
promoter, making its expression arabinose inducible. This method never worked as the 
wrong plasmids were sent. Following this attempt, BAC modification was postponed to 
begin the search for a more suitable BAC transgene. 
C57BL/6J BAC Library Screen 

The RPCI-23 BAC library was screened since our attempts at BAC modification 
were not working well. The RPCI library had three major advantages to the Research 
Genetics library. First, the average insert size for this library is 200 kb, increasing the 
chances of obtaining a BAC that is suitable for use as an imprinted transgene without 
additional modification. Secondly, this library is the template for the mouse genome 
sequencing effort, and so complete BAC sequences could possibly be posted as the 
project continued. Thirdly, if a complete BAC sequence was desired for a given BAC 
from this library, it could be queued for sequencing. Similarly to the strategy with the 
129/Sv library, the filters were probed using both the 5' and the 3' Snrpn probes. 
Several Snrpn transgenes were identified from this library. They are listed as follows: 



37 



73N19, 78014, 148C8, 160G24, 227M21, 284G21, 359M8, 365C3, 382D19, 438B21, 
396A12, 421K3, 9715, 215A9, 264P24, 276M1 1, 425D18, 380J10, 391B10, and 437G6. 

After streaking for single colonies and verifying that the clones represented the 
Snrpn region, intact BAC molecules were subjected to restriction digestion with a variety 
of rare cutting enzymes, including Sail, Sacll, and Notl and resolved the resulting 
fragments by pulse field electrophoresis. The gels were transferred to nylon membranes 
and probed with various probes in the Snrpn area to produce a rough restriction map of 
each BAC and to determine whether it contained enough genetic information to serve as a 
putative imprinted transgene. BAC end sequences for many of these clones were readily 
available on the National Center for Biotechnology Informatics (NCBI) website, and so 
primers could be ordered and used to create probes for the end of each BAC. The 
following probes were used in determining BAC tiling pathways and restriction map: a 1 
kb Sacl-Notl fragment 5' of Snrpn, a 2.2kb EcoRl-EcoRV fragment that is 5 kb 3' of 
Snrpn, an oligo that localized to an interstitial A particle (IAP) that is located in the 5' 
end of the Ipw gene and more than 100 kb 3' of Snrpn, and an oligo that is 
complimentary to a small nucleolar RNA (snoRNA) that was known to exist 
approximately 65 kb 3' of Snrpn. Using these probes, BACs 215A9, 425D18, and 
380J10 were determined to be suitable for further analysis. BAC 215A9 was shown to 
extend 120 kb 5' of Snrpn and 20 kb 3' of Snrpn, BAC 425D18 spanned a region that 
was first predicted to cover sequence 40 kb 5' of Snrpn through 65 kb 3' of Snrpn, but 
was later discovered to extend 90 kb 5' of Snrpn. Finally, BAC 380J10 was shown to 
contain only 8 kb 5' of Snrpn, but extended nearly 140 kb 3' of Snrpn, into the IAP 
element and the start of the Ipw gene (Fig 3-3). 



Lambda Red Mediated Homologous Recombination 

Although intact BAC molecules that seemed well suited to be imprinted without 
further modification were available, it was desirable to create a subtle mutation in the 
BAC that would allow transgenic Snrpn expression to be distinguished from the 
endogenous message. A third method of BAC modification that was developed by 
Daiguan Yu and Don Court at the National Cancer Institute (NCI) was pursued. This 
method employed a defective lambda prophage in which the genes allowing entry into 
the lytic phase of the phage lifecycle were removed, leaving only the minimal elements 
required to express the red exo, beta, and gam genes under a temperature-sensitive 
promoter. 73 The first version of this system was a modified strain, in which the defective 
lambda prophage was integrated into the bacterial genome. Unfortunately, this version 
had some remaining phage genes present that would cause lysis of the bacterial cell when 
induced, since these genes were also initiated from the temperature-sensitive promoter. 
The other major problem was that this phage system was developed in an E. coli host 
strain that had the opposite dam methylation status to the BAC host strain. A BAC clone 
isolated from a dam methylation-positive strain was therefore methylated and could not 
transform into the strain bearing the lambda prophage. As a result, none of the 
aforementioned BAC clones were successfully transformed into this strain. Two 
improved versions of the same system were later produced that were based on the BAC 
host strain of E. coli, DH10B. One version featured the defective lambda prophage as an 
integrant into the DH10B genome, while the other harbored a defective prophage that 
was maintained as a stable episome in the bacterial cell. While the transgenic mice in 



this study were derived unsing unmodified BAC clones, the BACs have since been 
modified using the lambda Red system and will be the focus of future work. 
Production Of Transgenic Mice 

While continuing with efforts to modify the BACs, it seemed prudent to begin 
producing transgenic mice carrying unmodified Snrpn BACs. BAC DNA was prepared 
by large-scale plasmid preparation, followed by cesium chloride banding of the resulting 
DNA. Both nicked and supercoiled bands were pulled and used for production of 
transgenic mice. It was important to have extremely fresh BAC, as the larger RPCI 
BACs degraded rapidly after isolation, as evidenced by low yield of supercoils and a 
diffuse appearance on pulse-field gels. The BAC DNA was dialyzed against TE buffer 
and diluted for injection with water to a concentration of 2.5 ng/u,l. Transgenic mice 
were made by pronuclear injection into oocytes derived from superovulated, fertilized, 
FVB/NJ female mice. The injected oocytes were then implanted into pseudopregnant 
(B6D2) F, mice. When the resulting pups reached 3 weeks of age, they were tail clipped 
to obtain DNA and earpunched for identification. Genomic DNA was prepared from 
each tail piece and was screened for the presence of the transgene by PCR using both end 
sequences from the corresponding BAC. Three transgenic founder animals were 
produced with the 215A9 transgene, two lines were made with the 425D18 transgene, 
and five lines carried the 380J10 transgene. 

The transgenic founders were mated with wild-type FVB/NJ females and the 
resulting litters were genotyped at 3 weeks of age to make sure that the transgene had 
integrated into the germline of the founders (i.e. that the founders were not chimeric), and 
to establish additional transgenic animals to secure the survival of the transgenic line. 



40 

One transgenic mouse derived from the 380J10 BAC did not show germline transmission 
to subsequent generations, and most likely represented a chimeric mouse in which the 
transgene had integrated into the genome sometime after the first cleavage following 
fertilization. Once the remaining lines were established, the individual lines were 
characterized for intactness of the transgene and copy number (Fig. 3-4). 
Characterization Of Transgenic Lines 

Characterization of the transgenic lines was necessary to determine the number of 
BAC molecules that were integrated into the transgene locus as well as the integrity of 
the molecules that had integrated. DNA used to make transgenic mice tends to integrate 
into the mouse genome in tandem arrays. For the purposes of our study, it was 
imperative that we obtain some single copy transgenic lines, since we were adhering to 
the most stringent determination of the minimum amount of sequence information that 
was sufficient to confer proper imprinted expression. It was also important that the 
transgene be intact, since proper imprinted expression may be a result of the relative 
location of the pertinent elements. Furthermore, the BAC molecules had to be linear 
prior to integration into the genome, and since we injected supercoils and nicked BAC 
molecules, the location of the site of breakage was random. This increased the chances 
of deriving a non-intact transgenic line. 

Copy number was determined by preparing genomic DNA from mouse tails 
harboring the transgene, performing Southern blot analysis on DNA that was cut with a 
specific restriction enzyme, and comparing bands from two different probes that would 
yield bands of similar size upon the same restriction digestion. One probe was located on 
the transgene, while the other was not. The intensities of the bands were compared and 



41 

the copy number was estimated as either single copy or multi-copy. One of the 215A9 
lines, one of the 425D18 lines, and one of the 380J10 lines contained multiple copies of 
the transgene, leaving us with two 215A9 transgenic lines, one 425D18 transgenic line, 
and three 380J10 transgenic lines that represented single copy transgene integrations (Fig. 
3-4). 

Similarly, rearrangements were determined by creating a panel of DNAs from the 
different transgenic lines digested by a variety of restriction enzymes. These panels were 
used to produce Southern blots, which were probed with several probes from the region 
(Fig 3-5). If a probe illuminated a band that was of a different size than that of a wild- 
type, non-transgenic mouse, then that transgene was considered rearranged. One 215A9 
line and one 380J10 line appeared to be rearranged by this analysis. However, due to the 
relatively few non-repetitive probes in the sequence surrounding the Snrpn locus, we may 
not have detected other rearrangements. 
The Marked Endogenous Snrpn Locus 

The inability to mark the BACs prior to their injection into mouse oocytes forced 
the use of a previously engineered Snrpn deletion mouse strain (ASmN) 12 . This mouse 
strain carries a deletion of exons 4-7 of the Snrpn gene, that results in almost complete 
absence of the Snrpn gene, and the low-level production of a larger fusion transcript that 
includes the first exons of Snrpn and the neomycin resistance gene. Using this allele, the 
endogenous Snrpn locus could be marked, alleviating the absolute requirement for a 
marked transgene (Fig 3-6). 

The APWS-IC mouse strain was also used. 22 This strain has a neomycin 
resistance cassette inserted in place of exons 1-5 of Snrpn. When paternally inherited, 



this deletion prevents expression from the Snrpn locus, allowing us to measure Snrpn 
expression that originates from the transgene (Fig 3-6). 
Imprinted Expression From Transgenes 

Since the transgenic copy of Snrpn was not subtly marked to be easily 
distinguishable from the endogenous allele, the ability of the transgene to show proper 
imprinted expression had to be tested by a complex breeding scheme using a marked 
endogenous allele. This breeding scheme is shown in Figure 3-7. First, the expression 
upon maternal inheritance was examined. Females who harbored a transgene were mated 
with a male mouse that carried the SmN deletion allele. The resulting pups were 
genotyped for the presence of the transgene, as well as for the presence of the deletion 
allele. Pups with both the transgene and deletion allele were sacrificed and their brains 
were used to make RNA. The RNA was analyzed by northern analysis for the expression 
of Snrpn. This is possible because the transgenic female mice transmit a transgene, 
which may or may not be expressed, depending on its epigenetic state, as well as a wild- 
type copy of the endogenous Snrpn locus that is not expressed since it is inherited from a 
female. The male mouse, on the other hand transmits the ASmN allele only, which is 
expressed as a larger, less abundant message. Any wild-type Snrpn message, would then 
come only from the transgene. Only the 380 transgenes were able to express Snrpn upon 
maternal inheritance, indicating that this transgene did not contain the sequence 
information that was required for silencing of the Snrpn gene upon maternal inheritance 
(Fig 3-8). 

Next, expression upon paternal inheritance was tested. In this case, male mice 
from the above described mating who had inherited both the transgene as well as the 



43 

ASmN allele were mated to females who were homozygous for the ASmN allele. The 
resulting pups were sacrificed and genotyped for the presence of both the transgene and 
the ASmN allele. The rationale behind this scheme is that the female in this case can only 
send a ASmN allele, which is silenced. In the presence of a known maternal ASmN 
allele, a paternally inherited wild-type allele can be distinguished from a paternally 
inherited ASmN allele. The male mouse transmits the transgene, which may or may not 
be expressed depending on its epigenetic state, and either a wild-type or ASmN allele. If 
the male transmits the wild-type allele, the resulting pups cannot be used for analysis, 
since the paternally inherited Snrpn locus is indistinguishable from the transgenic Snrpn 
locus. Alternatively, if the male transmits the ASmN allele, the endogenous allele that is 
expressed produces the longer, less abundant ASmN message (Fig 3-6). 

Originally, the idea was to identify Snrpn mice that were transgenic and 
homozygous for the ASmN allele, and examine expression in these pups. However, the 
presence of a large transgene that produced a wild-type Snrpn allele by Southern blot 
hindered identification of ASmN homozygotes. Instead, the mice were genotyped for the 
transgene, and RNA was made from the brains of the resulting transgenic mice. The 
RNA was then used to program RT-PCR reactions. PCR was performed on the RT 
product, using primers for both the ASmN and wild-type alleles. Pups that express both 
the ASmN and wild-type alleles inherit the ASmN allele and not the endogenous wild- 
type allele from their father (Snrpn and the ASmN alleles are only paternally expressed). 
Thus the wild-type allele must be expressed from the transgene. If the pup did not show 
expression from the ASmN allele, then it was believed to have inherited the wild-type 
endogenous allele from its father, and was uninformative as far as transgene expression is 



44 

concerned. From this assay, lines from each the two transgenes, 215A9 and 425D18, 
were shown to have expression upon paternal inheritance. Thus both 215A9 and 425D18 
are capable of showing correct imprinted expression in single copy transgenes (Fig 3-9). 

These experiments were repeated using the APWS-IC deletion mice in place of 
the ASmN deletion mice. To ascertain transgene expression upon maternal inheritance, a 
transgenic female mouse is mated with a APWS-IC male mouse (Fig 3-1 OA). The 
resulting pups were sacrificed and genotyped. Pups that inherited both the APWS-IC 
allele and the transgene were used to produce brain RNA, which was subjected to 
northern blot analysis and probed for Snrpn expression. Again, the females transmit the 
transgene as well as an epigenetically silenced maternal Snrpn allele, while APWS-IC 
males transmit a APWS-IC allele that completely lacks Snrpn expression. Any Snrpn 
message would have to originate from the transgenic copy of Snrpn. All five transgenic 
lines were subjected to expression analysis upon maternal inheritance, and again, only the 
380J10 transgenic lines, with the exception of line 380C, showed Snrpn expression upon 
maternal inheritance, verifying that they did not contain sufficient sequence to confer 
epigenetic silencing to the maternally inherited transgene (Fig 3-10B). 

To ascertain expression upon paternal inheritance, APWS-IC females were mated 
with transgenic males, and male pups with both the APWS-IC allele and the transgene 
were identified by genotyping. These pups survive since the APWS-IC allele is 
maternally inherited and therefore exists on the chromosome that is already 
epigenetically silenced. Male pups that have inherited both the APWS-IC allele as well 
as the transgenic allele were mated with wild-type females from the C57BL/6J strain to 
produce pups that harbor both the APWS-IC and the transgene (Fig. 3-1 1 A). These pups 



do not express endogenous Snrpn since their paternally inherited allele is the APWS-IC 
allele, which is silenced because the major promoter for the Snrpn gene is missing and 
because it is epigenetically silenced as a result of the imprinting center mutation. The 
only Snrpn expression would come from the transgenic Snrpn allele, and would occur 
only if the transgene were expressed upon paternal transmission. In this assay, the pups 
were sacrificed at birth and genotyped for the APWS-IC allele and the transgene, since 
pups that inherit the APWS-IC allele from their father suffer from the Prader-Willi 
phenotype and die shortly after birth. Pups that inherited the transgene as well as the 
APWS-IC allele were used to make brain RNA. This RNA was subjected to northern 
analysis, using a probe for the Snrpn message (Fig 3-1 IB). Two transgenic lines— 215B 
and 425 A— were identified that expressed transgenic Snrpn upon paternal transmission. 
Since neither of these lines were expressed upon maternal transmission, they were 
determined to demonstrate correct imprinted expression. 

Line 425A contains a single copy transgene that is intact. The transgene in this 
line is estimated to include the 22 kb Sail fragment that hosts the entire Snrpn gene, as 
well as 90 kb 5' of Snrpn and 65 kb 3' of Snrpn. This establishes the minimal genomic 
sequence that is sufficient to allow for correct imprinted expression in this region to be 
approximately 170 kb. Line 215B, on the other hand, is a single copy transgene that has 
undergone both 5' and 3' truncation upon integration into the mouse genome. The 5' 
truncation occurs at least 38 kb 5' of the Snrpn gene, while the 3' truncation occurs 
approximately 5 kb 3' of Snrpn. While this data suggests that the minimal sufficient 
region for correct imprinting is approximately 65 kb, the specific rearrangements that led 
to the truncation of this transgene are not known. The transgene in this particular line 



46 

may have also sustained additional mutations that influence its ability to be correctly 
imprinted. 

Discussion 

The 85 kb PI clone that shows correct imprinted expression when present in 
multiple copies, but not when present in single copy was later found to only be 65 kb in 
length. The discrepancy in length occurred because the PI was isolated from the 129/Ola 
mouse strain, which was found to have an insertion of a 20 kb VL-30 element into the 
region 5' of Snrpn. The parent-of-origin independent expression of this PI clone, now 
known to be 65 kb in length, indicates that either the PI clone is missing important 
regulatory sequences that lie 5' or 3' of the Snrpn gene, or the PI clone does not contain 
enough buffer sequence surrounding the necessary regulatory regions to establish a 
domain of imprinted expression. The PI clone is expressed upon both paternal and 
maternal inheritance when present in single copy, indicating that the defect is most likely 
in the ability to be silenced in the maternal germline. Since the murine AS-IC has not 
been functionally or physically identified in mouse and is postulated to be the negative 
regulatory element required for the silencing of the positively acting PWS-IC (Fig 2-4), 
the most likely explanation for the absence of maternal silencing of the PI transgene is 
that it does not contain the AS-IC regulatory sequences. Absense of the AS-IC would be 
predicted to prevent silencing of the positive element located on the transgene upon 
maternal inheritance. Since the PI clone without the VL-30 insertion extends 23 kb 5' of 
Snrpn, we believe that the AS-IC is located further than 23 kb 5' of Snrpn. This data 
spurred the search for transgenes with additional sequence 5' to Snrpn, rather than seek 
additional buffer sequence. Whether it is indeed the absence of regulatory elements or 



buffer sequence is a germane argument in this system, since we were seeking minimal 
sufficient sequence to achieve appropriate imprinting, rather than determining which 
specific elements are necessary. The data from the PI clone suggested that the 65 kb 
sequence was not sufficient to confer appropriate imprinted expression, and a single-copy 
imprinted transgene is necessary to establish the minimal sequence that can accomplish 
correct parent-of-origin dependent expression. Once the minimal required sequence has 
been determined, modifications can be made to that sequence to delineate specific 
elements that are necessary for the establishment and maintenance of the imprint. 

While initial attempts to make transgenic mice with the 320 kb YAC clone were 
unsuccessful, the YAC clone was used to establish the approximate distance between the 
3' most exon of Snrpn and exons B and C of Ipw to be lOOkb. However, the difficulties 
surrounding purification of a YAC of that size for transgenic injection precluded its 
convenient use as an imprinted transgene. Previously, the largest YACs that were 
successfully used in gene regulation studies involving transgenic mice were only 250 kb 
in length. 

Screening the 129/Sv BAC library proved fruitful for mining additional sequence 
5' of Snrpn, but did not provide an ideal BAC clone to use as an imprinted transgene. 
Two clones were obtained, 397F16 and 573H8 that together contained additional 
sequence 5' of Snrpn as well as a complete Snrpn gene. Unfortunately, the commercially 
available BAC libraries were of the 129/Sv or C57BL/6J strains, which made finding 
naturally expressed polymorphisms unlikely since both strains are from the domesticus 
subspecies of Mus musculus. The possibility of using a more polymorphic human Snrpn 



48 

transgene was explored and rejected earlier, since the human transgene is not silenced 
when present as a murine transgene regardless of copy number. 70 

Since the 129/Sv BACs were not ideal, the first strategy was to join the two BACs 
to create a larger B AC that was more suitable for transgene imprinting. The strategy was 
to place a neomycin resistance cassette into a 35kb Sail fragment derived from the 
397F16 BAC that lies 5' of Snrpn, and then excise this fragment and ligate it into the 
573H8 BAC near the site of exon 1 of Snrpn. At the time, only the RecA mediated 
homologous recombination method had been described. This method involved a two- 
step gene replacement strategy in which the RecA protein, absent in the BAC host strain, 
was provided from a pSVl -derived plasmid with a temperature-sensitive origin of 
replication 71 . Attempts to subclone homology arms into the RecA-containing plasmid 
failed, probably due to the large size of the plasmid. RecET mediated homologous 
recombination was also attempted. This method involved inducing RecE and RecT from 
an arabinose-inducible promoter. 72 This method was never fully explored because the 
wrong plasmids were sent. 

In the absence of a reliable homologous recombination method for BACs, the 
C57BL/6J BAC library that was available from the RPCI was screened. The BACs from 
this library were on average larger than those from the 129/Sv library. Additionally, the 
NCBI mouse genome sequencing effort was proceeding from this library, so the specific 
sequence in the region would eventually become available, as well as the addresses of 
additional clones in the region. Screening efforts for this library were very successful, as 
20 clones were isolated with the Snrpn probes. From the 20 clones, two were chosen that 
contained the most detectable sequence upstream of Snrpn while still carrying the entire 



49 

gene. While neither of the two BACs were in the sequencing program, both BACs were 
present as T7 and Sp6 fingerprints (end sequences) in the NCBI database. 

Pulse-field gel analysis initially indicated that BAC 425D18 contained 40 kb of 
sequence 5' of Snrpn, while 215A12 contained nearly 120 kb 5' of Snrpn. These 
estimates are rough, since they are based on large pulse-field fragments. Subsequent 
analysis suggested that 425D18 actually extended 80 kb 5' of Snrpn, and that the 40 kb 
Sail fragment that was originally thought to be the junction fragment between the mouse 
DNA insert and the BAC plasmid backbone was a doublet of two adjacent fragments. 

Three lines of transgenic mice harboring BAC 215A9 and two lines harboring 
425D18 were obtained from microinjection of mouse oocytes. These lines were named 
215A through C and 425 A and B. Among these five lines, lines 215B, 215C, and 425 A 
were all suspected to be present in single copy. Lines 215A and 425B were estimated to 
be present in 2-4 copy arrays. The method of analysis of transgene copy number that was 
used is probably accurate in delineating single-copy transgenes, but not in determining 
the copy number in multi-copy lines. Additionally, all transgenic lines except 215B 
appeared to be intact. 

Transgenes were first tested for correct imprinted expression using the ASmN 
strain of mouse. Expression upon maternal inheritance was readily assayed using the 
breeding scheme in Figure 3-7. Lines 215 A-C and lines 425 A and B were not expressed 
upon maternal transmission, suggesting that they could be silenced in the maternal 
germline (Fig 3-8). Lines 380A,B, and D, however were not correctly silenced upon 
maternal inheritance (Fig. 3-8) and were thought to be missing the AS-IC. This would 
result in the failure to silence the positively acting PWS-IC and lead to expression upon 



50 

maternal transmission. Paternal transmission was not conclusively tested using the 
ASmN breeding scheme (Fig 3-7). Mice homozygous for the ASmN allele and 
heterozygous for the transgene could not be distinguished from the ASmN heterozygotes 
by southern blot. Instead, RNA was made from the brains of transgenic mice and 
subjected to RT-PCR. If the ASmN allele was present, it was assumed that the paternally 
inherited allele was the ASmN allele and not the wild-type allele, leaving any wild-type 
Snrpn expression to be from the paternally inherited transgene. According to this assay, 
lines 215C, 425 A, and 425B were all expressed upon paternal inheritance (Fig 3-9). This 
breeding scheme was not further pursued, since it required the production of RNA and 
subsequent RT-PCR from every transgenic mouse. It was also not ideal that the genotype 
of the mice was never certain with respect to the Snrpn versus the ASmN allele. 
Furthermore, the potential for leaky expression from the maternally inherited allele 
complicated an assay that required RT-PCR. 

The imprinting assays were again tested using the APWS-IC strain and the 
breeding schemes shown in Figures 3-10A and 3-1 1A. Snrpn expression upon maternal 
inheritance was tested in a similar manner and produced the same results as previously 
(Fig. 3-10B). Snrpn expression upon paternal inheritance of transgenes proved to be 
much simpler using the APWS-IC strain. Mice inheriting the APWS-IC allele were easily 
genotyped by PCR that was unimpeded by the transgene. These mice could also be 
identified shortly after birth by their failure to thrive phenotype. Expression from the 
transgene in pups with a paternally inherited transgene and a paternally inherited APWS- 
IC allele could then be assayed for Snrpn expression by northern blot analysis, which 
does not detect the leaky maternal expression. By northern analysis, any Snrpn 



51 

expression originated from the transgene. Using the APWS-IC strain, lines 425A and 
215B exhibited Snrpn expression upon paternal transmission of the transgene (Fig. 3- 
1 IB). Since both of these lines are also silenced upon maternal transmission, they are 
both correctly imprinted. 

Line 215C was first shown to exhibit Snrpn expression upon paternal inheritance 
using the ASmN strain, but not with the APWS-IC strain. The most likely explanation for 
this is that the transgene is indeed expressed upon paternal inheritance, but at a low-level 
that is detectable by RT-PCR, but not by northern analysis. It was precisely this 
possibility that led to the use of the APWS-IC strain to test transgene imprinting. It was 
desirable to maintain a high level of stringency in considering a transgene correctly 
imprinted, the possibility that low-level expression could come from a correctly 
imprinted repressed allele existed. Since the expression of Snrpn from line 215C could 
possibly represent leaky expression of a repressed allele and not appropriate imprinting, 
line 215C was not considered to be correctly imprinted. 

Lines 425 A and 215B demonstrated appropriate imprinted expression, indicating 
that they contained sufficient sequence to confer proper imprinting. This was an 
important finding, since the AS-IC has to date not been defined functionally or physically 
in mouse. According to the current model for imprinting in the PWS/AS region, 57 a 
correctly imprinted transgene must contain the minimal sequence to harbor the positively 
acting PWS-IC element as well as the negatively acting AS-IC element. The 
identification of a correctly imprinted transgene indicates that the AS-IC can be 
functionally defined in mouse, since in this instance the silencing function occurs on a 
transgene that is integrated into a random genomic location. The transgene then serves to 



52 

define an approximate location of the AS-IC, and combined transgene data suggests that 
it lies upstream of the PWS-IC, as it does in humans. The 425A line definitively narrows 
this boundary to a region within 90 kb of the Snrpn gene. The 215B transgenic line has 
undergone truncations at both the 5' and 3' end of the BAC, but is still imprinted. These 
truncations narrow the probable location of the AS-IC to a region 38 kb 5' of Snrpn, but 
since the nature of the transgene rearrangements are not well understood, the use of this 
data to suggest the location of the AS-IC must be done with caution. 

The ability to use the lambda Red system of homologous recombination in B ACs 
will greatly simplify the study of transgene imprinting. A latent mark can now be placed 
into the Snrpn locus, and its expression upon both maternal and paternal inheritance can 
be ascertained in a single generation. This same system can be used to make deletions in 
the most 5' end of BAC 425D18 to better identify the AS-IC. Then, specific elements 
can be deleted, replaced, or rearranged to test different models of imprinting mechanisms 
in the PWS/AS region. 



53 



5' probe 



Snrpn 
gene 

U4 



3' probe 



Ipw 
IAP 



23 kb 



22 kb| 



30 kb 



967 PI (murine) 
181J3 



397F16 



518H22 



573B8 



YAC 



Figure 3-1. Restriction map of YAC clone and 129/Sv BAC library clones. 

YAC and BACs were subjected to restriction digestion with rare cutting 
enzymes and analyzed by pulse-field gel electrophoresis and southern analysis. 
Thin vertical lines represent Sail sites. Black boxes represent the Snrpn gene or 
Ipw IAP element. The 5' probe is a 1 kb Sacl-NotI fragment, while the 3' probe 
is a 2.2 kb EcoRl-EcoRY fragment. 



54 



397F16 



Homologous recombination 



397-neo 



Sail digestion 



Ligation into Sail digested BAC 



573B8 



Snrpn 



Snrpn 



Cre recombinase 



Snrpn 



35 kb 



55 kb 



Figure 3-2. Strategy to combine BACs 396F16 and 573B8. The strategy for 
joining two BACs to create one of an ideal size is diagrammed. Vertical hash 
marks indicate Sail sites. Dashed line represents BAC vector. 



55 



5' probe 



Snrpn 
gene 



3' probe 



967 PI (murine) 



23 kb 



22 kb| 



30 kb 



380J10 



8kb 



140 kb 



425D18 



90 kb 



65 kb 



170 kb 



20 kb 



215A9 



Figure 3-3. Restiction map of C57BL/6 BAC clones. BAC clones were 
subjected to restriction digestion and analyzed by pulse-field electrophoresis and 
southern analysis. The 5' probe is a 1 kb Sacl-Notl fragment, while the 3' probe is 
a 2.2 kb EcoRl-EcoRV fragment. Vertical hash marks represent Sail sites. Not all 
clones are diagrammed. 



PI -967 



1707 - 1 copy, intact 
1737 - 2 copies, intact 



215 



A - 2-4 copies, intact 
B - 1 copy, rearranged 
C - 1 copy, intact 



425 



A - 1 copy, intact 
B - 2-4 copies, intact 



380 



A - 2-4 copies, intact 
B - 1 copy, intact 
C - 1 copy, truncated 
D - 1 copy, intact 



Figure 3-4. Transgenic lines investigated for imprinted expression. 

Copy number and intactness is indicated according to available probes. 



Snrpn locus 



D 



1 2 3 456789 10 

i 1 1 Mil l I I 



Figure 3-5. Location of probes around the Snrpn locus. Probe A represents 
A 0.8 kb PCR fragment 38 kb 5' of Snrpn. Probe B is a 1 kb Sacl-Notl 
fragment that is 30 kb 5' of Snrpn. Probe C is a probe for exon 1 of Snrpn. 
Probe D is a 2.2 kb EcoRl-EcoRV fragment 5 kb 3' of Snrpn. 



Snrpn locus 

1 2 3 456789 10 

1 — I — I M ill I I 



ASmN allele 

1 2 3 4 •< — I 8 9 10 

\\— I — I — I hOEH-* 



APWS-IC allele 

< — I 8 9 10 

Neo H — H 



Figure 3-6. Snrpn locus, ASmN allele, and APWS-IC allele structures. 

The ASmN allele represents a deletion of exons 5-7 of the Snrpn gene. The 
Neomycin resistance cassete (Neo) produces a larger fusion transcript by 
Northern blot, thus marking that allele. The APWS-IC allele represents a 
35 kb deletion that includes 16 kb of sequence 5' of Snrpn as well as exons 
1-7 of the Snrpn gene. 



Maternal transmission 



Paternal transmission 



I 

^ P k ° 

ko 



Figure 3-7. Breeding scheme for testing expression upon maternal and 
paternal transmission of transgenes. Tg represents any transgene, while 
ko represents the ASmN allele. Maternal and paternal alleles are indicated 
where relevant. Only offspring of the desired genotype are indicated. 



60 




Figure 3-8. Expression of Snrpn from transgenes upon maternal transmission. 

The upper band results from a paternally inherited ASmN allele that creates a 
fusion transcript that is larger than the normal Snrpn message. Snrpn is only 
expressed from transgenes upon maternal inheritance in lines 380A, 380B, 
and 380D. 



61 




Figure 3-9. Expression of Snrpn from transgenes upon paternal inheritance. 

A. PCR primer locations for each allele. B. RT-PCR showing ASmN and 
wild-type alleles. The presence of the ASmN allele indicates that the 
paternally inherited endogenous allele was the ASmN allele, and then 
expression of the wild-type allele indicates expression of the transgene. If 
the ASmN allele is not present, then the paternally inherited endogenous allele 
is assumed to be a wild-type allele and this mouse is uninformative for transgene 
expression. 



62 



A 




Figure 3-10. Transgenic expression of Snrpn upon maternal transmission 
of transgenes. A. Breeding scheme used to assay transgenic expression using 
the APWS-IC strain . Tg represents any transgene, while ko represents the 
APWS-IC allele. Maternal and paternal alleles are indicated where relevant. 
Note that APWS-IC pups are missing all paternally expressed gene products 
in this region and succumb to neonatal lethality when the APWS-IC allele is 
paternally inherited. B. Northern blot showing Snrpn expression in transgenic 
pups inheriting a paternal APWS-IC allele and a maternal transgenic allele. 



63 



1st 

Generation 




2nd 
Generation 





X Cf Tg ^ m 




„ + m 




T « P to P 



B 



Snrpn 




Figure 3-11. Transgenic expression of Snrpn upon paternal transmission 
of transgenes. A. Breeding scheme used to assay transgenic expression using 
the APWS-IC strain . Tg represents any transgene, while ko represents the 
APWS-IC allele. Maternal and paternal alleles are indicated where relevant. 
Note that APWS-IC pups are missing all paternally expressed gene products 
in this region and succumb to neonatal lethality when the APWS-IC allele is 
paternally inherited. B. Northern blot showing that Snrpn is expressed upon 
paternal expression in lines 425 A and 215B. 380 lines are not shown 



CHAPTER 4 

STRAIN-DEPENDENT DIFFERENCES IN PHENOTYPE 

Introduction 

Angelman Syndrome (AS) and Prader-Willi Syndrome (PWS) are caused by 
deficiencies in genes subject to genomic imprinting. PWS is characterized by infantile 
hypotonia, gonadal hypoplasia, short stature, a moderate delay in physical and mental 
development, and obsessive/compulsive behavior, as well as neonatal feeding difficulties 
followed later by hyperphagia leading to profound obesity. 74 AS is characterized by 
severe mental retardation, absent speech, ataxic gait and a happy demeanor. 24 
Approximately 70% of PWS patients have a 3-4 megabase deletion of the paternal 
chromosome 15ql 1-ql 3. 49,75,76 The clinically distinct AS results from the same 15ql 1- 
ql3 deletion in about 70% of patients, however the deletion is always on the maternally 
inherited chromosome. 45,77,78 Either syndrome can also result from uniparental disomy 
(UPD), in which both copies of chromosome 15 are inherited from only one parent. The 
UPD is always maternal in PWS patients 28,49,76 and paternal in AS patients. 31,79 The 
identification of both of these classes of patients has led to the conclusion that PWS is 
caused by a loss of gene expression from the paternally inherited chromosome, whereas 
AS is caused by a loss of gene expression from the maternally inherited chromosome. 

Many AS patients have been shown to contain intragenic mutations in the UBE3A 
gene, indicating that mutations in UBE3A are sufficient to cause AS. 34,35 In contrast to 
AS, no single gene has been identified for PWS, strongly suggesting that PWS is a 



64 



contiguous gene syndrome, requiring the loss of two or more paternally expressed genes 
to cause the PWS phenotype. 80 All the PWS candidate loci identified to date in humans 
and mice are expressed exclusively from the paternally inherited chromosome. These 
include: MKRN3/Mkrn3 (previously known as ZNF127IZfpl27), 4681 MAGEL2IMagel2, 
42.65 NDN/Ndn> 40 SNRPN/Snrpn, M ' 82 HBII-486, HBII-13/MbII-13, 43 HBII-487, HBII- 
438A, HBII-85/MbII-85, 4344 IPW/Ipw, 1847 HBII-52/MbII-52, 43 HBII-438B, and an 
antisense transcript to the UBE3A gene. 41,57 Recently, the latter genes (SNRPN through 
the antisense UBE3A transcript) were shown to be derived from a single transcriptional 
unit. 83 

The imprinted genes involved in AS and PWS are regulated by a bipartite 
imprinting center (IC) located upstream of the SNRPN gene. 26 The IC is divided into the 
Angelman Syndrome Imprinting Center (AS-IC) and the Prader-Willi Syndrome 
Imprinting Center (PWS-IC). Previously, it was demonstrated that the PWS-IC is 
functionally conserved in the mouse. 22 A 35 kb deletion mutation (the deletion was 
originally reported to be 42 kb, but is now known that the original map was based on a 
strain that contained a VL30 insertion) was created in a male embryonic stem (ES) cell 
line. This 35 kb deletion included 16 kb of upstream sequence plus exons 1-6 of the 
Snrpn gene. Breeding male chimeric founders (129/Sv) harboring a maternal deletion 
mutation to wild-type females (C57BL/6J) resulted in mutant offspring that usually died 
within 48 hours, but never survived beyond 7 days after birth. These mice exhibited 
several phenotypes similar to those found in PWS infants including small size, poor 
feeding, and failure to thrive. In addition, these progeny were shown to lack expression of 



66 

the local paternally expressed genes Mkrn3 faka Zfpl27), Ndn, and lpv? 2 as well as 
Magel2, MbII-13, MbII-85, and MbII-52. 43,44 ' 84 

Since germline transmission from chimeric males produced mutant offspring 
that died prior to weaning, this 35 kb deletion mutation was recreated using J4 female 
(XO) ES cells. As expected, transmission of the deletion mutation from female chimeras 
resulted in normal heterozygous carrier offspring. 55 Consistent with previous 
observations, heterozygous carrier males (129/Sv) bred to C57BL/6J females produced 
mutant offspring that did not survive until weaning, even if the wild-type siblings were 
removed to reduce competition. This strain has been used to breed unaffected males 
APWS-IC carriers with mice of different genetic backgrounds to study phenotypic 
variation. Here, long-term survival of APWS-IC mice upon breeding to several strain 
backgrounds is reported. 

Materials and Methods 

Strains and Matings. 

Mouse strains used were APWS-IC deletion mice backcrossed onto a C57BL/6J 
genetic background for 10 generations, C57BL/6J, FVB/NJ, 129/Sv, C3H/HeJ, DBA/2J, 
and C57BL/6J congenic for a region of Mus musculus castaneus chromosome 7 
(B6.CAST.c7). 56 APWS-IC males were mated with females of the strains described 
above to produce affected PWS progeny pups, which were used to ascertain survival and 
gene expression. For (B6.CAST.c7 x FVB/NJ) F, matings, B6.CAST.c7 females were 
mated with FVB/NJ males. Conversely, (FVB/NJ x B6.CAST.c7) F, matings were 
produced using FVB/NJ females and B6.CAST.c7 males. 



67 

Culling and Fostering 

Litters resulting from matings with APWS-IC males were selectively culled to aid 
survival of affected APWS-IC pups. Culling was carried out by sacrificing all of the 
wild-type pups except one within two days after birth, depending on whether or not wild- 
type pups could be distinguished from their APWS-IC littermates. One wild-type pup 
was left to stimulate milk production in the absence of robust pups. Mice were fostered 
to another mother by removing the APWS-IC pups from one mother (usually from the 
C57BL/6J strain) and giving them to the other mother (usually from the FVB/NJ strain) 
whose pups had been removed. Additionally, urine from the foster mother was wiped on 
the foster pups to aid in her acceptance of the new pups. In some fostering cases, two 
mothers of different strains were set up to litter together, and the litter from the FVB/NJ 
mother was then removed and the wild-type littermates from the other mother were 
culled. This allowed two new mothers to care for a small number of APWS-IC pups. 
Identification of Polymorphisms 

Total brain RNA was prepared from both 129/Sv mice and B6.CAST.c7 mice. 
Each RNA sample was used to create single- stranded cDNA using random primers in a 
reverse transcription reaction. These cDNAs were used to program PCR reactions using 
primers corresponding to Ube3a exons 4 and 7 (5 ' -CCTGC AGACTTG AAG AAGC AG- 
3' and 5 ' -G AAAACCTCTGCG AAATGCCTT-3 ' ). The resulting products were cloned, 
sequenced and compared. A polymorphism found in exon 5 created a Tsp509l restriction 
site in the 129/Sv clone and a Bst4Cl site at the same location in the B6.CAST.c7 clone. 

A polymorphic Avail site between Mus musculus castaneus and Mus musculus 
domesticus was identified and kindly provided to us by Marisa Bartolomei (M. 



Bartolomei, personal communication). This polymorphism results from a cytosine 
nucleotide, rather than a thymine nucleotide at position 1 17 of the Ndn transcript. Avail 
cuts the domesticus allele, but not the castaneus allele due to the presence of the cytosine 
nucleotide. PCR primers flanking the polymorphic restriction site were designed as 
follows: NdnpolyF 5 '-ACAAAGTAAGGACCTGAGCGACC-3 ' and NdnpolyR 5'- 
C AAC ATCTTCT ATCCGTTCTTCG- 3 ' . The PCR product amplified using NdnpolyF 
and NdnpolyR was gel purified on a 2% agarose gel, extracted from the gel using the 
QIAGEN gel extraction kit, and digested with Avail. The digested products were 
electrophoresed on a 4.8% agarose gel (2:1 low-melt agarose: agarose). 
RT-PCR 

Total RNA was isolated from complete brains obtained from neonatal, 1 week 
old, 2 week old, and 3 week old mice. RNA was extracted using RNAzol (Tel-Test, Inc.) 
according to instructions. The total brain RNA (lOug) was pretreated with DNasel 
(Invitrogen), and half of the reaction was subsequently used to synthesize first-strand 
cDNA with Superscript II reverse transcriptase (RT) and random primers (Invitrogen). 
The other half of the reaction was manipulated in parallel in the absence of RT. One- 
twentieth of the +RT or -RT reactions was used to seed PCRs using the following 
conditions: lOmM Tris-HCl, pH 8.3, 50mM KC1, the four dNTPs at 0. 125 mM each, 1 
unit Taq DNA polymerase (Boehringer), and the appropriate primers at 4 uM each. 
Sequences of the primers were as follows: Ndn 9F (5'- GTATCCCAAATCCACA 
GTGC -3'), NdnlOR (5'- CTTCCTGTGCCAGTTGAAGT -3'), Ndnpoly F (5'- ACAAA 
GTAAGGACCTGAGCGACC -3'), Ndnpoly R (5'- CAACATCTTCTATCCGTTC 



TTCG -3'), Ube3a 5F (5'- CACATATGATGAAGCTACGA -3'), Ube3a intron 5 R (5'- 
CAGAAAGAGAAGTGAGGTTG -3'), Smpn Nl.l (5'- CTGAGGAGTGATTGCA 
ACGC -3'), Snrpn N2.2 (5'- GTTCTAGGAATTATGAGCCCC -3'), p-actin F (5'- 
GTGGGCCGCTCTAGGCACCAA - 3'), and P-actin R (5'- CTCTTTGATGTCA 
CGCACGATTTC - 3'). PCR amplifications conditions were 95° C for 5 min., followed 
by either 20 or 30 cycles of 94° C for 30 s, 60° C for 45 s, and 72° C for 45 s. The final 
cycle was followed by an extension step for 10 min at 72° C. 
Northern Blot Analysis 

To detect the presence of processed snoRNAs, total RNA was separated on 8% 
denaturing polyacrylamide gels (7M urea, 1 X TBE buffer) and transferred to nylon 
membranes (Hybond N+) using a semi-dry blotting apparatus (7>arc.s-blot SD, BioRad) as 
described by Cavaille et al. 43 RNA was fixed to the membrane by baking at 80° C in a 
vacuum oven. The membranes were prehybridized and hybridized overnight at 58° C, 
according to the Church and Gilbert method. Oligonucleotides complementary to 
snoRNA sequences were end-labeled using a 32 P-dATP (NEN) and T4 polynucleotide 
kinase (Invitrogen). Their sequences are as follows: MbII-85 (5' - TTCCGATGAGAG 
TGGCGGTACAGA - 3'), MbII-52 (5' - CCTCAGCGTAATCCTATTGAGCATGAA - 
3'), and 5.8S rRNA (5' - TCCTGCAATTCACATTAATTCTCGCAGCTAGC - 3'). 
Probes were purified from free nucleotides using the nucleotide removal kit (QIAGEN). 
Membranes used for detecting snoRNA expression were washed twice for 15 min at 
room temperature using 2X SSC, 0.1% SDS, and then exposed to Kodak XAR film at - 
80° C for 12-36 hours. 



70 

Results 

In contrast to the fully penetrant neonatal lethality phenotype previously seen on 
the C57BL/6J strain background, APWS-IC carrier males on the 129/Sv strain 
background or APWS-IC carrier males made congenic on the C57BL/6J background bred 
to wild-type female FVB/NJ mice, produced several mutant offspring that survived to 
two weeks of age. Furthermore, if most of the wild-type competitor sibs are removed, 
the mutant offspring could survive to adulthood. Although these surviving APWS-IC 
mice are significantly smaller than wild-type sibs (Fig. 4-1), both males and females are 
fertile. 

To examine whether the increased survival was simply due to superior mothering 
abilities of the FVB/NJ females as compared to C57BL/6J females, newborn pups were 
fostered to mothers of the opposite strain and also trio matings of a APWS-IC carrier 
male with both an FVB/NJ and a C57BL/6J female were established. In both situations, 
survival of only those F, mice born to the FVB/NJ females was observed, demonstrating 
that it is the genetic background of either the offspring or the mother that leads to survival 
rather than the skill of the mother (data not shown). 

Next, the expression of several PWS candidate genes in these surviving (FVB/NJ 
x APWS-IC) Fi mice was examined. Surprisingly, low levels of Ndn and antisense Ube3a 
expression was detected by RT-PCR in newborn mice (data not shown). By Northern 
blot, expression of several snoRNAs mapping to the region was also detected (data not 
shown). These results were unexpected as expression of these genes in pups with a 
paternally inherited PWS-IC deletion had not been previously detected. However, when 
RNA expression from the 5' half of the Snrpn gene, a region that is physically deleted 



71 

from the paternally inherited APWS-IC allele, was examined, expression was detected by 
RT-PCR (data not shown). Therefore, this strongly suggested that low-level expression 
detected for Ndn, antisense Ube3a, MbII-85 and Snrpn is derived at least somewhat from 
leaky expression of the maternal FVB/NJ allele. A time course for the observed low- 
level expression of Ndn and Snrpn at newborn, 1 week, 2 weeks, and 3 weeks of age was 
followed, and the leaky expression was detectable in all ages of mice (Fig4-2). It was 
hypothesized that expression of these genes is sufficient to rescue the lethality phenotype 
associated with the APWS-IC mutation. 

Next, the extended survival and leaky expression of these imprinted genes was 
investigated to determine if it was restricted to the FVB/NJ strain. Females for the 
C57BL/6J, DBA/2J, BALB/cJ, C3H/HeJ, and 129/SvEv strains were mated to APWS-IC 
carrier males on the C57BL/6J genetic background. Wild-type pups were removed from 
the litters within two days after birth and the longest possible survival of APWS-IC pups 
was determined. Pups derived from C57BL/6J and DBA/2J females had the shortest 
survival time, with the longest surviving affected individual from either strain living only 
7 days. Pups derived from all other strains had at least one affected individual that 
survived for at least 3 weeks. Leaky expression of Snrpn, Ndn, and MbII-85 was 
observed at comparable levels in all strains (Fig. 4-3), regardless of their ability to 
survive. 

Finally, the leaky maternal expression of these PWS candidate genes was 
investigated in wild-type mice. FVB/NJ females were mated with wild-type males from 
the B6.CAST.c7 strain, a strain that is congenic for the mus musculus castaneus 
chromosome 7 on a C57BL/6J background. The reciprocal mating was also performed. 



Total brain RNA was isolated from the resulting (FVB/NJ X B6.CAST c.7) F, or 
(B6.CAST.c7 X FVB/NJ) F, mice at birth. RT-PCR followed by a polymorphic 
restriction enzyme digest of the RT-PCR product was used to distinguish between 
expression of Ndn produced from the domesticus and castaneus alleles. Only the paternal 
castaneus allele was expressed in each case (Fig 4-4). B6.CAST.c7 females were then 
mated with APWS-IC carrier males and the same RT-PCR and restriction digestion on 
brain RNA made from (B6.CAST.c7 X APWS-IC) F, pups was performed. While 
expression from wild-type pups was limited strictly to the domesticus allele, the 
expression in pups that had inherited the paternal APWS-IC allele was biallelic, 
representing both domesticus and castaneus alleles. This indicates that leaky maternal 
expression is only observed or detectable when in combination with a paternal APWS-IC 
mutation. 

Discussion 

This study demonstrated that strain background affects the phenotype of the 
APWS-IC mice. Heterozygous carrier males (either 129/Sv or C57BL/6J) bred to 
C57BL/6J or DBA/2J females produced mutant offspring that do not survive until 
weaning, even if wild-type siblings are removed to reduce competition. However, long- 
term survival was achieved upon breeding these same carrier males to several other 
strains, including FVB/NJ, C3H/HeJ, 129/Sv, and BALB/cJ. While surviving APWS-IC 
mice are smaller than wild-type littermates, they do not become obese on a low-fat diet, 
both males and females are fertile, and they appear to be normal in all other respects. 

Expression analysis of Snrpn, the antisense transcript of Ube3a, and the MbII-85 
snoRNA reveals leaky expression of these paternally expressed genes in APWS-IC pups. 



73 

This expression occurs at a very low level, and originates, in part, from the maternally 
derived chromosome. This leaky expression, originally detected in PWS-IC deletion 
pups with FVB/NJ mothers, is also present at comparable levels in all other strains, 
including C57BL/6J and DBA/2J. This suggests that the increased survival of APWS-IC 
pups is not due to the low-level expression of genes in the PWS region. Furthermore, this 
expression is not sufficient to ameliorate the failure to thrive phenotype and permit 
survival in APWS-IC pups. Similarly, humans that have PWS due to maternal 
uniparental disomy or a balanced translocation have also been reported to show weak, but 
detectable expression of genes in the PWS region. 85 No obvious clinical differences were 
reported from individuals with leaky expression from imprinted genes when compared to 
affected individuals that did not show leaky gene expression from these genes. 85,86 

Why would APWS-IC pups survive on some strains and not others, when the gene 
expression in the PWS region appears to be identical between survivors and non- 
survivors? The most likely explanation for this would be the presence of specific 
modifier alleles in some inbred strains that either permit or prevent survival of affected 
pups. An alternate explanation may be the general unfitness of a particular strain for 
survival of neonates. This is unlikely, however, as the APWS-IC pups are genetically F, 
individuals and should benefit from hybrid vigor. 

Heterozygous carrier males on the 129/Sv genetic background bred to C57BL/6J 
females, produce APWS-IC pups that never survive. Conversely, carrier males on the 
C57BL/6J genetic background bred with 129/Sv females produce APWS-IC pups that are 
capable of surviving to adulthood, even though the genetic background of these 
reciprocal matings are identical. One explanation of this is that the modifier gene or 



74 

genes responsible for survival of APWS-IC mice could be required in the oocyte. The 
oocyte specific factor would not be present in C57BL/6J oocytes that are not permissive 
for survival, but would be present in 129/Sv oocytes and allow survival of affected pups. 
An alternate explanation is that the modifier gene or genes are themselves imprinted. If 
the modifier gene was only expressed from the maternally inherited chromosome, and the 
allele of this gene that was permissive for survival was the 129/Sv allele and not the 
C57BL/6 allele, then survival could only be achieved when the maternal allele of this 
gene is from the 129/Sv strain. Pronuclear transplant, where the pronuclei from a (B6 X 
APWS-IC) F, oocyte could be moved into "emptied" oocytes from fertilized FVB/NJ or 
(FVB/NJ X APWS-IC) F, females, could be used to distinguish between these two 
possibilities. If the transplanted pronuclei gave rise to pups that were able to survive, 
then the modifier gene is most likely oocyte specific, however, if the resultant pups still 
do not survive, then the modifying gene is most likely imprinted. 

It is interesting that the leakiness of the maternal allele is only detectable when in 
combination with a paternal APWS-IC allele, but not when in combination with a wild- 
type paternal allele. There are three possible explanations for this. First, low-level 
expression from the maternal allele would produce a few transcripts from the maternal 
allele that could simply be swamped by the more abundant transcripts originating from 
the paternal allele. Perhaps imprinting in mouse is only strong allele bias, and low-level 
expression could mark the repressed chromosome as a region that is repressed in a 
facultative manner as opposed to a constitutive manner. Another possibility is that the 
expression of genes or transcripts downstream of Snrpn could be allowed by the splicing 
of the upstream exons of Snrpn 87 that are left intact on the APWS-IC chromosome into 



75 

the downstream exons of Snrpn or the run-on transcripts from which the snoRNAs are 
spliced. 83 These upstream exons have been shown to be functional alternate promoters 
for Snurf, but their level of expression is much weaker than that of the major promoter. 
Perhaps this could explain the weak expression of the snoRNAs. Additionally, the 
physical interaction between the silenced imprinting center and the upstream genes, 
including Mkrn3, Magel2, and Ndn, could be required for the silencing of these genes 
that lie more than 1 Mb away from the imprinting center. The missing imprinting center 
could cause the alternate Snurf promoters to be juxtaposed to the regulatory elements of 
the upstream genes, and result in leaky expression. While this may explain leaky paternal 
expression, it doesn't explain residual expression from the maternal chromosome. The 
third explanation is that communication between one silenced and one expressed 
imprinting center is required for the proper silencing of the maternal allele and the 
expression of the paternal allele. Specifically, the unmethylated, unsilenced, paternal 
PWS-IC may be required in trans to silence the maternal region. While this explanation 
could account for the bi-allelic expression in the APWS-IC mice, it doesn't explain why 
the expression levels of the imprinted genes are so low. Furthermore, frans-activation is 
a phenomenon not documented in mouse and would be difficult to prove or disprove. 
Leaky gene expression in the region is not unique to mice, as human patients with 
maternal uniparental disomy have also been reported to show leaky expression 85,86 . 




Figure 4-1. Surviving APWS-IC mice are smaller than their wild-type 
littermates. APWS-IC mice and their wild-type littermates were weighed at 3, 6, 9, 
12, and 15 weeks of age. Their weight in grams is plotted versus their age in weeks. 
Filled symbols represent APWS-IC mice and open symbols represent the wild-type 
littermates. 



77 



Ndn RT-PCR 



5'Snrpn RT-PCR 



P-actin RT-PCR 



No RT control 



MbII-85 




MbII-52 



5.8S rRNA 




Figure 4-2. Leaky expression from paternally expressed genes is not age 
dependent. APWS-IC mice were sacrificed at birth, one week, two weeks, and 
three weeks of age. Total brain RNA was made from the sacrificed mice and 
subjected to either RT-PCR or northern analysis. RT-PCR is shown for Ndn and 5' 
Snrpn since the expression level is not detectable by northern analysis. 5.8S rRNA 
and P-actin were used as controls for northern analysis and RT-PCR, respectively. 
Low-level expression is noted in APWS-IC mice regardless of their age. 



78 



B6 FVB/N C3H/HeJ Balb/c 129/Sv DBA/2J 



Ndn RT-PCR 



5'Sn/p/i RT-PCR 



P-actin RT-PCR 



No RT control 



Ndn RT-PCR 

20 cycles, blot 



S'Snrpn RT-PCR 
20 cycles, blot 



MbIl-85 



w A AwAAwAAvvA Aw A A w A A 









- - - 


• • • 


















_3 





# © o o • 

* 

tuiii 




5.8S rRNA 



Figure 4-3. Leaky gene expression does not account for strain dependent 
differences in phenotype. APWS-IC mice made congenic on the C57BL/6J strain 
background were mated with either C57BL/6J, FVB/NJ, C3H/HeJ, Balb/cJ, 
DBA/2J, or 129/Sv females. Total RNA was prepared from the brains of two 
APWS-IC mice and one wild-type littermate. The RNA was subjected to RT-PCR 
and northern analysis. Ndn and S'Snrpn were analyzed by RT-PCR at 40 cycles or 
at 20 cycles and blotted and probed. 5.8S rRNA and fi-actin were used as controls 
for northern analysis and RT-PCR, respectively. APWS-IC mice showed leaky 
expression in all strains analyzed. 



79 



FVB/NJ x B6. Cast.c7/B6.Cast.c7 x FVB/NJ 



FxC CxF 
genomic genomic 



castaneus 
domesticus 




B6.Castc7 x APWS-IC 



castaneus 
domesticus 




Figure 4-4. Leaky gene expression is biallelic in APWS-IC pups. A. Brain 
RNA from reciprocal crosses between FVB/NJ and B6.CAST.c7 was subjected 
to RT-PCR with primers flanking a polymorphic Avail site in the Ndn transcript 
from mus musculus domesticus. The product was digested with Avail and run 
on a 4.8% agarose gel. Wild-type Ndn expression in both crosses originates 
from the paternally inherited chromosome. B. B6.CAST.c7 females were 
mated with males that were carriers for the APWS-IC allele. APWS-IC and wild- 
type littermate pups were sacrificed at birth and total RNA was prepared from 
their brains. RT-PCR and Avail digestion of the Ndn transcript was performed 
as described above. Wild-type Ndn expression occurs from the paternally 
inherited allele, while APWS-IC expression occurs from both paternal and 
maternal alleles. 



CHAPTER 5 

TRANSGENIC RESCUE OF THE PWS-IC DELETION MOUSE 

Introduction 

Prader-Willi syndrome (PWS) patients first present with severe infantile 
hypotonia and failure to thrive that leads to a requirement for gavage feeding in most 
PWS neonates. The infantile phenotype subsides after several months and gives way to a 
period of normal feeding behavior, but soon progresses to hyperphagia and severe 
obesity, if uncontrolled. 23 PWS children also possess a distinctive behavioral phenotype 
that is similar, but not identical to obsessive-compulsive disorder. Outwardly visible 
signs of the disease in humans may also include almond-shaped eyes, strabismus, small 
stature, small hands and feet, and hypogonadism. 23 ' 37 Caused by the physical or 
functional deletion of paternal chromosome 15ql l-ql3, PWS occurs at a frequency of 
about 1/15,000 live births. 24 ' 26 Approximately 70% of PWS patients have sustained a 
large (3-4 Mb) deletion of their paternally inherited chromosome 15. Other PWS patients 
have either inherited both copies of chromosome 15 from their mother (maternal 
uniparental disomy), or have an imprinting mutation. Patients with an imprinting 
mutation have either a deletion in the 5' region of Snrpn, including exon 1, an area 
known as the imprinting center (IC) of the 15ql l-ql3 region, 611 " 13,20 or have no detectable 
DNA sequence mutation, but still possess a paternally inherited chromosome that 
epigenetically behaves as a maternally inherited chromosome. 



80 



81 

Human chromosome 15ql l-ql3 contains at least six paternally expressed genes 
or transcripts and is syntenic with central chromosome 7 in mouse. The genes and 
transcripts include: MKRN3, MAGEL2, NDN, SNRPN, HBII-13, HBII-436, HBII-437, 
HB1I-438A, HBII-438B, HBII-85, HBII-52, and the t/#E3A-antisense transcript. 1718,21,39 ' 44 
Expression of each of these genes is completely absent in most PWS patients, although 
some patients show low-level expression, but still present with classical PWS. 85,86 A 
large region of sequence between SNRPN and NDN still remains unexplored and may 
contain additional genes. This region encompasses nearly 1 .5 Mb, and additional genes 
discovered in this region might be predicted to be imprinted and paternally expressed. 

The distinct paternal inheritance pattern of PWS identifies 15ql 1-ql 3, as well as 
the syntenic mouse chromosome 7 as a region of genomic imprinting. 17,21,39 Genomic 
imprinting involves the expression of only one parental allele in somatic cells. The other 
parental allele is transcriptionally silenced. The non-transcribed alleles of the PWS 
genes, maternal in the case of PWS-related genes, are often associated with nearby 
methylated cytosine residues present in CpG islands, while the transcribed alleles, are 
associated with the lack of methylation at these same CpG residues. 611 " 13 This indicates 
that the maternal allele has a different epigenotype than the paternal allele. It is not 
known whether the eventual silencing of the maternal allele causes this epigenetic mark, 
or whether it is a mark to identify the allele to be silenced. 

Conversely, Angelman syndrome (AS) results from the functional deletion of the 
maternal complement of the same interval of chromosome 15. 24 A unique class of AS 
patients with specific mutations in the UBE3A gene suggests that loss of UBE3A 
expression can be solely responsible for the major clinical features of AS. 34,35 Another 



maternally expressed gene, ATP10C, has also been identified in the PWS region, but its 
contribution to the phenotype of AS is yet to be determined. 36 AS can result from 
maternal deletion of 15ql l-ql3, paternal uniparental disomy, IC mutations, and from 
putative single gene mutations, as evidenced by completely normal epigenotype in the 
region and the UBE3A mutation class which present with classical AS. 24 The absence of 
the class of PWS patients in which a single gene mutation is responsible for the complete 
phenotype, as well as the presence of multiple paternally expressed genes that are 
disrupted, indicate that PWS is a contiguous gene syndrome where the clinical features 
are the cumulative effect of multiple disrupted genes. 80-88 

PWS and AS patients with imprinting defects have either a relatively small 
deletion or no noticeable mutation at all, but these mutations alter the entire 3-4 Mb 
region on the paternal and maternal chromosomes, respectively. 6-20 In PWS-IC mutation 
patients, the paternally derived genes assume maternal, or silenced, epigenotypes. The 
deletion breakpoints of several PWS patients with IC mutations were mapped to identify 
the smallest region of deletion overlap for PWS (PWS-SRO). 13-37 The PWS-SRO is 
located 5' of the SNRPN gene, includes the first exon of the gene, and is often referred to 
as the PWS-IC. Although the PWS-SRO has been narrowed down to approximately 
4.3kb, the actual deletions that have been shown to cause PWS are much larger. 37-38 A 
smallest region of deletion overlap has also been identified for AS, and is referred to as 
the AS-SRO. The AS-SRO lies around 35 kb upstream of the PWS-SRO, and has been 
narrowed to 0.8 kb. 37-38 AS-IC mutations never appear to overlap the PWS-SRO, while 
PWS-IC mutations can overlap the AS-SRO, suggesting that in the absence of the PWS- 
IC, the AS-IC is not necessary. 57 This indicates that the sole purpose of the AS-IC may 



83 

be to regulate the PWS-IC in the maternal germline. 57 This also indicates a bipartite 
structure to the IC, in which the PWS and AS components work together to control 
imprinting in the 15ql l-ql3 region. In mouse, the PWS-IC has been functionally defined 
by deletion, indicating that the bipartite IC is most likely functioning similarly between 
humans and mice. 22 

Our lab previously created a mouse model of PWS that emulates both the 
phenotypic and molecular characteristics of PWS by deletion of the mouse PWS-IC. 22 
Mice that inherit a PWS-IC deletion (APWS-IC) from their father lose expression of the 
known paternal-specific genes, including Snrpn, Mkrn3, Ndn, and the Ube3a antisense 
transcript. The APWS-IC mice also demonstrate poor suckle and failure to thrive, 
indicative of the infantile characteristics of human PWS. These mice thus provide an 
excellent mouse model of PWS. Unfortunately, APWS-IC mice on the inbred C57BL/6J 
background die within seven days of birth, presumably a result of poor suckle and failure 
to thrive. Progression to obesity and the development of other PWS related phenotypes 
has yet to be seen in APWS-IC mice due to neonatal lethality. APWS-IC mice mated 
with mothers of other strain backgrounds survive to adulthood if the wild-type siblings 
are removed shortly after birth. These surviving APWS-IC mice remain smaller than 
their wild-type siblings, but are otherwise grossly normal. Furthermore, surviving 
APWS-IC females and males are both fertile and never become obese on an ordinary 
rodent diet. 

Individual knockouts of Snurf, Snrpn, and Mkrn3 appear completely normal and 
show no PWS associated phenotype. 22 ' 46 89 " 91 Conflicting data regarding the phenotype of 
the Ndn gene knockout have been reported. Tsai and colleagues report that the Ndn 



knockout mouse has no discernable phenotype, while Gerard et al. and Muscatelli et al. 
report partial neonatal lethality that is strain dependent and not consistent from generation 
to generation. 89,90,92 On the other hand, mice inheriting a paternal deletion spanning from 
exon 2 of Snrpn to Ube3a succumb to neonatal lethality in 80% of the affected pups. 91 
Together, this data strengthens the hypothesis that PWS is a contiguous gene syndrome 
caused by the absence of multiple gene products, and not caused by the absence of any 
single gene product. Despite the efforts of many labs to create individual gene 
knockouts, the contribution of each gene in human chromosome 15ql l-ql3 and the 
syntenic mouse chromosome 7 to the PWS phenotype is still a mystery. Figure 5-1 
shows a summary of the individual gene deletions that have been reported. 

Efforts to model PWS using individual gene knockouts have been largely 
unfruitful. 22,89,90 ' 92 However, engineering multiple gene deletions is also limited by the 
location of the imprinting centers (several desired mutations would remove them, 
creating an imprinting mutation). Therefore, a different approach was adopted to 
determine the individual gene contribution to this contiguous gene syndrome. This 
approach involved transgenic rescue, and the strategy was to produce transgenic mice 
harboring either individual or multiple PWS genes and mate them to APWS-IC mice. In 
mice paternally inheriting an imprinting center mutation and maternally inheriting a 
transgene, the ability of the transgene to rescue the phenotype of the APWS-IC mice 
could be ascertained. The genes responsible for the individual aspects of the PWS 
phenotype could then be definitively determined, since one transgene could conceivably 
rescue the neonatal lethality defect, while another could rescue the small stature defect. 
In addition, the rescued mice may live longer or recapitulate other human PWS 



85 

symptoms and allow the dissection of additional later onset phenotypes in our mouse 
model. 

Materials and Methods 

Screening the BAC Libraries 

The Research Genetics (Huntsville, AL) BAC library was initially screened for 
BAC clones. This library was made from the 129/Sv strain of mouse. The library 
consisted of 9 filters, each carrying 27,648 unique clones, with an average insert size of 
130 kb, spotted in duplicate onto 9 different membranes. Probes for Ndn, Mkrn3, Ube3a, 
and the gene formerly known as Ipw were used to screen the library simultaneously. 

Prior to hybridization, BAC membranes were washed with 1500 mL of 6X SSC 
and 0.1% SDS at room temperature for 15 minutes. The membranes were then rinsed 
twice with 1500 ml of 6X SSC for 15 minutes at room temperature, and hybridized 
according to the Research Genetics protocol. Briefly, the membranes were prehybridized 
in roller bottles with 3 membranes per bottle, each separated by one sheet of Flow Mesh 
(Diversfied Biotech, Boston, MA), in 120 ml of HyperHyb (Research Genetics, 
Huntsville, AL) per bottle at 65° C for 20 minutes. At least 10 7 counts of each probe 
were boiled for 5 min, snap cooled on ice, and then added to 3 ml of HyperHyb that was 
pre- warmed to 65° C, and 1 ml of hybridization solution was added to each of the roller 
bottles. The membranes were allowed to hybridize for 2 hours at 65° C. 

The membranes were washed in the roller bottles three times for 15 minutes at 
65° C with 30 ml IX SSC and 0.1% SDS. They were then removed from the bottles and 
washed twice more at 65° C with 1000ml of IX SSC and 0.1% SDS for 15 minutes. 
Finally, the membranes were rinsed with IX SSC at room temperature. The membranes 



86 

were then wrapped in cellophane and exposed to film (XAR, Kodak) overnight at -80° C. 
Positive clones were identified following the manufacturer's instructions, and the 
following clones were ordered: 454N20, 205L17, 581E10, 173C16, and 143C10. In 
addition to screening the 129/Sv library, a single RPCI-23 BAC, 452P17 was found by 
screening the library virtually with a probe to the 3' most intron of the Ube3a gene. 
Probe Preparation 

Probes were labeled using the Prime-It II kit (Stratagene, La Jolla, CA), with 
minor modifications. DNA (200 ng)was boiled for 5 minutes in the presence of random 
hexamers. The DNA-hexamer mix was cooled on ice, and 5X reaction buffer was added 
along with 25 uCi of a 32 P dCTP and 1U of exo" klenow. The labeling reaction was 
placed at 37° C for 15 minutes. After labeling was complete, the probe was purified 
using the nucleotide removal kit (QIAGEN) and eluted in 200 ul of water. Purified 
probes were boiled for 5 minutes and snap cooled on ice prior to use in hybridization. 
Verification of BAC Clones 

Single BAC colonies were tested to determine which clones were derived from 
each gene locus. Four single colonies from each BAC were streaked onto LB plates 
supplemented with the antibiotic, chloramphenicol (LB-CA). The plates were overlayed 
with a nylon filter and grown overnight at 37° C. The next morning, the bacteria were 
lysed on the filters as follows: first, the filters were soaked in alkali solution for 4 
minutes, followed by neutralizing solution for 4 minutes, and finally rinsed in 6X SSC for 
1 minute. The filters were then baked at 80° C for 2 hours, pre-hybridized in Church and 
Gilbert buffer for one hour at 65° C, and hybridized with the Ndn probe. Upon washing 
and laying the filters to film, the results were used to determine which clones were 



87 

derived from the Ndn locus and whether the culture sent represented a pure colony from a 
single clone, or a mixed culture with multiple BAC clones. The filters were then stripped 
in boiling 0.5% SDS and the entire hybridization procedure was performed again with the 
remaining probes. 
End Sequencing BAC Clones 

The BAC clones were end sequenced using the big dye termination reaction to 
obtain sequences unique to the mouse genomic DNA insert. Reactions were performed 
on 1 mg of BAC DNA that had been sheared by passing it five times through a 25 gauge 
needle, using a modified version of the manufacturer's instructions, with 12 ul of 
sequencing buffer, 4 \xl of big dye, 16pmol of primer, and 2.5mM MgCl 2 . The reactions 
were subjected to an initial 5 min. cycle at 95° C, followed by 30 cycles of 95° C for 30 s, 
50° C for 10 s, and 60° C for 4 min. The completed reactions were purified from free 
nucleotides using Performa DTR Gel Filtration Columns (Edge Biosystems), vacuum 
dried, and sequenced by the University of Florida Center for Mammalian Genetics 
Sequencing Core. 
Pulse-Field Gel Electrophoresis 

Pulse-field gel electorphoresis was carried out to create a rough restriction map 
and tiling pathway for the BAC clones. BAC DNAs digested with the appropriate 
enzymes were electrophoresed on a 1% genetic technology grade agarose gel (Nusieve, 
FMC) made with IX TBE buffer. Samples were mixed with 10X ficoll loading dye and 
ran at 200V with a switch time of 1-12 seconds for 12.5 hours. This provided resolution 
of bands from 6-200 kb. The gel was stained with a dilute ethidium bromide solution, 
photographed, and blotted as described below. 



88 

Southern Blot 

Agarose gels carrying 5 \ig of digested genomic DNA or 1 u.g of digested BAC 
DNA were photographed and UV nicked for 5 minutes. The gel was soaked in Alkali 
solution for 45 minutes, followed by soaking in neutralizing solution for 90 minutes. The 
gel was transferred to Hybond nylon membrane (Amersham) in 10X SSC overnight. The 
membrane was rinsed in 2X SSC and baked at 80° C for 2 hours. Hybridization was 
carried out according to the method of Church and Gilbert. Membranes were 
prehybridized in 20 ml of Church and Gilbert Buffer for 1 hour at 65° C. The 
prehybridization solution was poured off and replaced with 5 ml of fresh solution and the 
boiled, cooled probe. Hybridization continued overnight at 65° C. Washing was carried 
out three times in 2XSSC and 0.1% SDS at 65° C for 15 minutes each. 
Transgenic Mice Production 

For production of transgenic mice, BAC DNA was prepared by large scale 
plasmid preparation, using the alkaline lysis method. First, log phase cultures of BAC 
culture grown in LB-CA were centrifuged for 5 minutes at 6,000 rpm to pellet the cells. 
The supernatant was poured off and the pellet was resuspended in GTE solution plus 
lysozyme. The suspension was then lysed with NaOH and SDS. After addition of the 
lysis solution, the lysate was neutralized in 0.5M potassium acetate, and poured through 
cotton gauze into Oak Ridge tubes. BAC DNA was precipitated with ethanol, and the 
pellet was resuspended in 4 ml TE with 10 u.g/ml RNAse A, and allowed to incubate at 
37° C for 1 hour. Cesium chloride and ethidium bromide were added to the DNA 
solution, and the solution was loaded into a heat sealable tube. The tube was sealed and 
spun at 58,000 rpm for 18 hours. BAC supercoils were viewed with a UV light, and both 



nicked and supercoiled bands were pulled. The DNA was extracted with tert-butanol 
until the ethidium bromide was completely depleted from the BAC DNA. The DNA was 
then dialyzed twice against 1 L TE buffer, using dialysis tubing for 6 hours each. The 
concentration of the BAC DNA was then measured by spectrophotometer, and then 
diluted to 2.5 ng/u.1 for microinjection. 

Transgenic mice were made by pronuclear injection into oocytes that were 
obtained from superovulated, fertilized, FVB/NJ female mice. The fertilized, injected 
oocytes were then implanted into (B6D2) F, females that were made pseudopregnant by 
mating them with vasectomized (B6D2) F, males. 
Mouse Husbandry 

At 3 weeks of age approximately 2mm of tail was clipped from each mouse, and 
the ear was punched for identification. Genomic DNA was prepared from the tail piece 
by incubation in tail lysis buffer with proteinase K (100 ng/u.1) overnight at 55° C. The 
tail lysate was extracted with phenol: chloroform: isoamyl alcohol (25:24:1) and then 
precipitated with ethanol. The tail DNA was screened for the presence of the transgene 
by PCR using both end sequences from the BAC. 
Mouse Strains 

APWS-IC mice were developed in our lab 22,55 and constituted a 35 kb deletion 
encompassing the sequence from 16 kb 5' of Snrpn and 19 kb 3' of Snrpn. These mice 
were backcrossed for 10 generations to C57BL/6J males and were thus made congenic on 
that strain background. 

Transgenic mouse strains were originally maintained on the FVB/NJ genetic 
background, but selected strains were then backcrossed for 10 generations to the 



90 

C57BL/6J strain background. Four strains of 454N20 transgenic mice are referred to as 
454A-D, one strain of mice representing the PI clone 967 is referred to as line 1707, and 
5 lines representing transgene 380J10 are referred to as 380A-E. 

Results 

Screening the BAC Library for Clones 

The first step toward transgenic rescue was to screen a commercially available 
BAC library for clones that would putatively contain the PWS genes. BAC clones were 
chosen to perform the rescue as they are likely to contain all of the necessary regulatory 
regions for the individual genes due to their size. They are also relatively easy to handle, 
as they are propagated in bacteria, rather than yeast. Additionally, they can be modified 
by homologous recombination. 71 " 73 The Research Genetics BAC library was first 
screened for clones. This library is made from the 129/SvEv strain of mouse, and the 
average insert size for BAC clones is 130kb. Some clones by screening the RPCI-23 
BAC library were also obtained. This library is made from a female C57BL/6J mouse. 
The average insert size for this library is 200 kb and is the template for the public mouse 
genome sequencing effort. 

The BAC libraries were screened using probes for Ndn, Mkrn3, and the gene 
formerly known as Ipw. With the first screen of the Research Genetics library, two 
BACs were obtained that hybridized with the probe for Mkrn3, one of which also 
hybridized with the Ndn probe. No clones for Ipw were identified. From subsequent 
screenings of this library, one BAC containing Ipw, one that hybridized with a probe for 
the Ube3a gene, and one that hybridized with Ndn, Mkrn3, and Magel2 were isolated. 
From screening the RPCI library with the Snrpn probes, a single BAC that spanned most 



91 

of the distance between Snrpn and Ipw was isolated. An electronic database search also 
identified a single BAC that contained the 3' most intron to Ube3a. 

Addresses for the BAC clones were determined, the BACs were ordered, and then 
they were streaked out to determine their clonality and to verify the presence of the 
probes that the library was screened for. The span of the individual BAC clones were 
determined by end sequencing, making end probes by PCR, and using them to probe 
southern blots produced from restriction digestion with rare cutting enzymes, separated 
by pulse-field electrophoresis. From this it was determined that the previously published 
gene order in the Ndn region was incorrect. The correct order places Magel2 between 
Mkrn3 and Ndn, and in a transcriptional orientation opposite that of Mkrn3. The BACs 
that were isolated were 454N20, which contains Mkrn3, Magel2, and Ndn; 205L17, 
which contains Frat3 and Mkrn3; 173C16, which contains Ube3a; and 380J10, which 
extends from Snrpn to the MbII-85 snoRNA cluster (Fig 5-2). 
Production and Characterization of Transgenic Mice 

Once the BACs of interest had been identified and mapped, transgenic mice were 
made. BAC DNA was prepared by large scale plasmid preparation, followed by cesium 
chloride banding of the resulting DNA. Both nicked and supercoiled bands were pulled 
and used for production of transgenic mice. It was important to have extremely fresh 
BAC, as the larger BACs from Roswell Park degraded rapidly after isolation. The DNA 
was injected at 2.5 ng/ul. Transgenic mice were made by pronuclear injection into 
superovulated, fertilized, FVB/NJ female mice. The fertilized oocytes were then 
implanted into pseudopregnant (B6D2) F, mice. When the resulting pups reached 3 
weeks of age, they were tail clipped to obtain DNA and earpunched for identification. 



92 

The tail DNA was screened for the presence of the transgene by PCR using both end 
sequences from the BAC. Four transgenic founder lines were made with BAC 454N20, 
and five were made with BAC 380J10. 

The transgenic founders were mated to ascertain whether or not the transgene 
bred true. This indicates whether or not the founder was perhaps a chimera, and therefore 
may not have the transgene present in its germline. Only one line from the 380J10 BAC 
did not breed true. The established lines were then characterized for rearrangements and 
copy number. Rearrangements were determined by assembling a panel of DNA from 
transgenic mice digested with different restriction enzymes. The panel of cut DNAs was 
used to produce Southern blots which were then probed with various probes that were 
known to be on the transgene. A rearrangement was detected if a band present in a 
transgenic mouse was a different size than that of a wild-type mouse. Rearrangements 
were detected in lines 454B, 454C, and 380C. Copy number was determined crudely by 
comparing the intensity of a band represented by both endogenous alleles and the 
transgenic allele with a similarly sized band that is not located on the transgene. The 
copy number was simply estimated as single copy or multi-copy. One single copy and 
two multi-copy lines were obtained with the 454 transgene (the rearranged line was not 
tested). For the 380J10 transgene, two single copy and two multi-copy lines were 
obtained. Figure 5-3 summarizes the characterization of the transgenic lines with respect 
to the copy number and integrity of the BAC transgene within each line. 

Next, whether or not the individual genes were actually expressed from the 
transgene in each transgenic line was determined. This was difficult because a human 
transgene or a reporter construct was not used to distinguish transgenic expression from 



93 

endogenous expression. Instead, transgenic expression was analyzed by mating a female 
transgenic mouse with a APWS-IC male. In the pups inheriting a paternal APWS-IC 
allele, the paternally expressed genes are silent. If a APWS-IC male is mated with a 
transgenic female, any expression of the local paternally expressed genes would be from 
the transgene in pups that have both the APWS-IC allele and the transgene. The breeding 
scheme for obtaining APWS-IC deletion mice that are also transgenic is shown in Figure 
5-4. Expression was examined by northern blot analysis and by RT-PCR performed on 
brain RNA from pups that had inherited both the transgene and the APWS-IC allele. By 
northern blot, lines 454A and 454D show detectable Ndn expression from the transgene 
(Figure 5-5). Probes for Magel2 and Mkrn3 did not work in Northern analysis, so they 
were analyzed by RT-PCR. These genes were also expressed from the transgenes in the 
454A and 454D transgenic line. Transgenic expression from lines 454A and 454D was 
observed by RT-PCR for all three genes, while line 454B did not show expression of 
Mkrn3, Magel2, or Ndn. For the 380 transgenic lines, Snrpn was detected by northern 
analysis in 380A, 380B, and 380D (Fig 5-6). However, MbII-85, a snoRNA cluster that 
is present on the transgene was only detectable by northern in line 380A, indicating that 
there may be some undetected rearrangements in lines 380B and 380D (Fig 5-6). 

One caveat to the expression and rearrangement data is that quantitation of the 
expression level of each gene in the transgenic mice has only been ascertained in a rough 
fashion. Individual genes are under the control of their native promoters and regulatory 
elements since the BAC transgenes contain several kilobase pairs of flanking sequence. 
Rough estimates from northern analysis of the 454 transgenes show that Ndn seems to be 
expressed at near wild-type levels (Fig 5-5). In contrast, MbII-85, a snoRNA harbored on 



94 

the 3 80 A transgenic line is expressed at a reduced level compared to wild-type as 
analyzed by northern analysis (Fig 5-6). However, Snrpn expression appears to occur at 
close to wild-type levels in all 380 and 1707 transgenic lines (Fig 5-6). Precise 
expression level beyond that described has not seemed pertinent. 

The rearrangement detection capabilities in the BAC transgenic mouse lines are 
also limited. Sequences upstream of Snrpn are riddled with LINE elements and local 
repeats that have precluded development of probes near the endpoints of the 454 BAC. 
Similarly, the repetitive nature of the snoRNA clusters precludes development of useful 
probes, and so rearrangements or breakpoints in BACs occurring 3' of Snrpn are nearly 
impossible to detect. 

Transgenic Rescue Experiments: 454 Transgenic Lines 

Transgenic female mice representing each 454 line were mated with males that 
were heterozygous for the APWS-IC allele (Fig 5-3). When pups were born, they were 
counted each day up until weaning, and dead mice and body parts were collected as they 
became available. The dead pups as well as the live pups were genotyped for the 
transgene as well as the APWS-IC allele. In the case of each transgenic line assayed, no 
pup survived to weaning that inherited the APWS-IC allele, even if they were also 
transgenic, indicating that the 454 transgenic lines, expressing Ndn, Magel2, and Mkrn3 
could not rescue the neonatal lethality phenotype. 

While mice with FVB/NJ mothers harboring a paternally inherited APWS-IC 
allele did not survive until weaning, they did frequently survive until 2 weeks of age. It 
was eventually determined that APWS-IC could survive to adulthood if the wild-type sibs 
were removed, and that surviving APWS-IC mice were smaller than their wild-type 



95 

littermates. While it was evident that the 454 transgenes could not completely rescue the 
neonatal lethality phenotype, it could be possible that the 454 transgenes could rescue the 
small stature phenotype observed in surviving APWS-IC mice. Again, 454 transgenic 
females were mated with APWS-IC carrier males, but then the wild-type littermates were 
removed at 2 days after birth, when they could be easily distinguished from the affected 
APWS-IC pups. The wild-type pups were fostered to a separate mother from our mouse 
colony so that they could be compared to the APWS-IC mice from the same litter. At 
weaning time (3 weeks of age), the pups were genotyped and weighed. The pups were 
again weighed once every 3 weeks until they were 15 weeks of age. Figure 5-7 shows a 
graph of age in weeks versus weight in grams. The 454A, 454B, and 454D lines were not 
capable of rescuing the small stature phenotype of APWS-IC mice. 
Transgenic Rescue Experiments: 380 and 1707 Transgenic Lines 

Similar to the 454 lines, female transgenic mice representing the three 380 lines as 
well as the single PI derived line were mated with APWS-IC carrier males. Pups were 
allowed to progress to weaning naturally, and dead bodies were collected. As with the 
454 transgenes, the 380 and 1707 transgenic lines did not permit survival of APWS-IC 
mice, suggesting that transgenes expressing Snrpn do not rescue the neonatal lethality 
observed in APWS-IC mice. 

The rescue mating was repeated and the wild-type pups were fostered to new 
mothers to remove wild-type competition. At 2 days after birth, all of the APWS-IC mice 
were readily identified, regardless of whether or not they carried a transgene. In 
transgenic lines 380B, 380D, and 1707, the transgenic littermates were not 
distinguishable by 3 weeks of age, either. Line 380C was not subjected to rescue 



96 

analysis, since Snrpn was not expressed in this line. In line 380A, however, while 
transgenic APWS-IC pups were indistinguishable from non-transgenic APWS-IC pups at 
birth, by 3 weeks of age the transgenic versus non-transgenic APWS-IC pups could be 
separated by size. The small pups were the APWS-IC pups, while the larger pups 
represented those that had inherited the 380A transgene as well as the APWS-IC allele. 
Pups from each transgenic line were weighed at weaning (3 weeks of age), and then 
every 3 weeks thereafter until 15 weeks of age. Figure 5-8 shows these growth charts for 
the 380 and 1707 transgenic lines. The 380A transgene appears to rescue the small stature 
phenotype of the surviving APWS-IC mice, but does not rescue the neonatal failure to 
thrive. Furthermore, wild-type mice that are transgenic for the 380A transgene are larger 
than their non-transgenic counterparts, suggesting that the 380A transgene promotes 
growth. None of the other transgenes rescue the small stature or the failure to thrive on 
the FVB/NJ strain background. 

Discussion 

BAC Isolation 

BAC clones proved to be ideal for use as transgenes in the transgenic rescue 
experiments for a variety of reasons. They are relatively easy to handle, being 
propagated and isolated like conventional plasmids. Intriguingly, the BAC clones that 
were isolated, 454N20 and 380J10, each contained multiple genes, facilitating the 
attempted transgenic rescue with the known paternally expressed genes. BAC clones 
containing each of the known paternally expressed genes were isolated, however, the 
BAC contig has not been extended across a 1.5 Mb region that lies 5' of Snrpn and 3' of 
Ndn. This region is highly repetitive, difficult to probe, and its size was underestimated 



97 

prior to the publication of the first draft of the public mouse genome sequencing project 
was done. However, the NCBI sequencing efforts have now identified a BAC fingerprint 
that spans this region. 

Transgenic mice produced from the BAC clones represented a mix of single-copy 
and multi-copy lines, with the overall efficiency of producing transgenic mice being 
about 1 in 10 live births (data not shown). The transgenes in this study were also 
transmitted in normal mendelian ratios (data not shown). Four transgenic lines with BAC 
454N20, four lines representing BAC 380J10, and a single line from the PI clone were 
produced. Lines 454A and 454D expressed Mkrn3, Magel2, and Ndn, line 454B did not 
express any of the three genes contained on the BAC, and line 454C was lost during 
maintenance of the strain, and was not analyzed further. Line 380A expressed Snrpn, 
MbII-85, and putatively all of the snoRNAs located between those two genes, while lines 
380B, 380D, and 1707 expressed only Snrpn. It is unknown whether or not these 
transgenic lines express any of the single copy snoRNAs, other than MbII-13, that lie 3' 
of Snrpn. 

Rearranged Transgenes 

Rearranged transgenes are inevitable when making transgenic mice using our 
technique, as BAC supercoils and nicked molecules from the cesium chloride banding 
procedure were injected. In the nicked fraction, the nick occurs randomly with no 
discernment against disruption of the mouse DNA insert. Similarly, the supercoils must 
break to integrate into the mouse genome, and so this break occurs randomly. The 
burden then falls on the ability to generate enough transgenic lines to obtain at least one 
that is not rearranged, and then on the ability to detect rearrangements. Since gene 



98 

expression was the assay for whether or not a transgene was functional, only a crude 
estimate of the locations of rearrangements was needed. It has also worked to our 
advantage that some of the rearrangements that were obtained unwittingly produce a 
deletion series for the rescuing BACs. For example, line 380A expresses Snrpn, MbII- 
85, and probably the snoRNAs that lie between them, while lines 380B and 380D only 
express Snrpn and some of the single-copy snoRNAs. This creates the opportunity to 
differentiate between the effects of Snrpn in a transgenic rescue paradigm and Snrpn plus 
MbII-85, without having to inject a different transgene. 

454N20 and 380J10 transgenic lines that do not express all of the genes that are 
located within their respective clones may do so for a variety of reasons. In the case of 
the 454B transgenic line, a rearrangement that disrupts or deletes the promoter region of 
Ndn is the most likely explanation for the inability of this line to express Ndn. 
Alternatively, either a regulatory region may be physically separated from the promoter 
region, resulting in disruption of Ndn transcription. Interestingly, it may be the case that 
Mkrn3, Magel2, and Ndn are regulated by a common enhancer in addition to the PWS- 
IC. In the case of the 380 transgenes, the Snrpn transcript has been postulated to drive 
expression of the heavily spliced transcript that the snoRNAs are spliced from, so 
disruption of the Snrpn promoter is not possible, since Snrpn expression seems normal in 
these lines. It is probable that the BAC clone was broken 3' of the Snrpn gene, although 
this rearrangement was not detectable with any of our probes. Development of additional 
probes in this region is difficult since it is highly repetitive. 

The expression of MbII-85 from the 380A transgene poses a curious observation. 
The expression level of this snoRNA is relatively low, compared to wild-type expression 



99 

levels. The 380A transgene may not contain the full MbII-85 cluster. The BAC itself 
may only extend into the first few copies of MbII-85, and the expression level seen by 
northern may be a result of the few copies present on the BAC. An alternate explanation 
is that the repeated transgene causes some silencing of expression from the trangene, 
although this is somewhat unlikely since Snrpn appears to be expressed at near wild-type 
levels. It is also formally possible that lines 380B and 380D do not have a rearrangement 
at all. They may express MbII-85 as well, albeit at a lower level. MbII-85 expression 
may be visible solely because of the multi-copy insertion of this transgene causing 
higher, more detectable expression of the MbII-85 snoRNA species. 
Transgenic Rescue 

That transgenic lines 454A and 454D do not rescue the neonatal lethality of the 
APWS-IC mouse is possibly very interesting in light of the Ndn knockout data. Two 
groups report neonatal lethality of variable penetrance in conjunction with a targeted 
mutation of the Ndn gene. 90,92 However, the 454 transgenes, which express Ndn, as well 
as Magel2 and Mkrn3, do not rescue neonatal lethality at all. While these data may seem 
at odds, it is important to remember that PWS is a contiguous gene syndrome, caused by 
the disruption of two or more paternally expressed genes. The inability to rescue the 
PWS phenotype with the Ndn- containing transgenes could be due to the presence of 
another unidentified gene that participates in the neonatal failure to thrive phenotype. 
Alternatively, the disparate results between the multiple labs creating Ndn knockouts may 
be revealing in themselves in that Ndn may not contribute significantly to failure to 
thrive, and that the Stewart and Muscatelli groups may be producing regulatory mutations 



100 

that affect genes other than Ndn. Alternately, there could still be differences in strain 
backgrounds that explain this. 

Nonetheless, it is important to point out that Ndn, Magel2, and Mkrn3 do not 
seem to have a major role in failure to thrive or in small stature. Transgenic lines 
expressing these genes do so at a level that is very close to wild-type levels and occurs 
from the native promoter, so lack of rescue is not likely due to incorrect expression levels 
or incorrect tissue specificity of transgenic expression. The most likely explanation is 
that there is more than one gene that contributes to the failure to thrive phenotype. If this 
is the case, we may have to complement with both genes in order to see rescue of the 
APWS-IC phenotype. Alternatively, these three genes may be involved in some of the 
behavioral phenotypes of PWS that we have not investigated in our mouse model. It is 
also possible that these genes are responsible for obesity in humans, and the rescue of an 
PWS-associated obesity phenotype cannot be investigated because the mouse model does 
not become obese. 

The ability of the 380A transgene to rescue the small stature phenotype of the 
APWS-IC mice is particularly intriguing. This rescue has first allowed the separation of 
the failure to thrive phenotype from the small stature seen in the APWS-IC mouse model. 
Secondly, it has attributed the small stature phenotype to the MbII-85 snoRNA cluster. 
Transgenic lines 380B, 380D, and 1707 each fail to rescue the small stature phenotype 
suggesting that neither Snurf nor Snrpn is responsible for the small stature phenotype. 
However, line 380A, the only line expressing MbII-85, does rescue this phenotype, 
suggesting that MbII-85 is at least somewhat responsible for growth. This is the first 
instance of a phenotype being specifically attributed to a snoRNA and assigns a function 



101 

to one of these illusive imprinted small RNAs in the region. It is also interesting that the 
expression level required to rescue this aspect of the APWS-IC mouse is relatively low. 

Many questions still remain regarding the transgenic rescue using the 380A BAC. 
First of all, if the expression level of MbII-85 were occurring at a wild-type level would 
the failure to thrive phenotype be ameliorated? Secondly, does 380A express the entire 
MbII-85 cluster? Subcategories of the MbII-85 species exist, and it is uncertain how 
many of these are included on the 380 BAC. Furthermore it is not known whether or not 
the different subcategories have different functions. However, these studies show that 
MbII-85 does play an important role in the PWS phenotype, indicating that the snoRNAs 
warrant further investigation. 

Remembering that PWS is a contiguous gene syndrome, matings between the 
3 80 A transgenic mice and 454D transgenic mice have been estabilished to obtain double 
heterozygotes that can be mated to APWS-IC males. These matings will yield a small 
percentage of mice that are transgenic for both 380A and 454D, but also inherit a paternal 
APWS-IC allele. In these mice, the ability of Ndn, Magel2, and Mkrn3 as well as Snrpn 
and MbII-85 to promote complete rescue of the PWS phenotype can be investigated. 
These transgenic mice can also be used to test the ability of the various transgenes to 
rescue the failure to thrive and neonatal lethality phenotypes on the C57BL/6J strain 
background, since five of the transgenic lines are being made congenic on this strain.. 



102 



/ 9 * 



A 
1 



A 

2-4 




► 

41 



AA 

5 6 



8,9 — 



10 



11 



12 

■//■ 



77^ 



77^ 



Snrpn 



Figure 5-1. Individual gene deletions in the PWS region. 

I . Mkrn3 targeted disruption shows no overt phenotype 46 . 

2-4. Ndn targeted disruption shows either neonatal lethality 89 or no overt phenotype 90 92 . 

5. Snurf targeted disruption shows no overt phenotype 91 . 

6. Snrpn targeted disruption shows no overt phenotype 22 . 

7. Deletion of the PWS-IC causes neonatal lethality 22 . 

8-9. Smaller deletions around the PWS-IC show incomplete IC deletion phenotype 87 . 
10. A deletion spanning from Snurf to Ube3a causes neonatal lethality and small 
stature 91 . 

II. A radiation induced mutation spanning from Ipw downstream causes no overt 
phenotype when paternally inherited (D. Johnson). 

12. A transgene insertion that causes a large deletion of the PWS/AS region causes a 
phenotype similar to the IC deletion mouse (R. Nicholls) . 



103 



/ / / * 



9 



9) 




4 



205A9 



Pl-967 



80 kb 
454N20 



75 kb 



452P17 



170 kb 



150 kb 



380J10 



170 kb 



Figure 5-2. Clones representing the PWS region. BACs or PI clones were isolated from 
129/Sv or C57BL/6J libraries. Their locations and sizes are approximated. 



454 



PI -967 



380 



104 



A - 2-4 copies, intact 
B - 1 copy, rearranged 
C - line lost 
D - 1 copy, intact 



1707 - 1 copy, intact 
1737 - 2 copies, intact 



A - 2-4 copies, intact 
B - 1 copy, intact 
C - 1 copy, truncated 
D - 1 copy, intact 



Figure 5-3. Transgenic lines obtained for transgenic rescue. Copy number and 
intactness is indicated according to available probes. 



105 




2 



Tg (FVB/NJ) 




1 



(5 



APWS-IC (C57/BL6J) 




wt 



Tg, wt 



APWS-IC 



Tg, APWS-IC 



Expresses all genes 



Expresses all genes 
plus genes from the 
transgene 



Expresses none of the 
genes 



Expresses none of the 
genes, except those 
provided by the transgene 



Figure 5-4. Breeding scheme for transgenic rescue. Sex and strain are indicated where 
relevant. 



106 



8 



Ndn 



I 



I 



oo 
en 



I 



IN 

8 



9 



I 




Figure 5-5. A^rfn, Magel2, and Mkrn3 are expressed in two transgenic lines. A. Northern 
blot analysis was used to investigate Ndn expression in mice that had inherited a paternal 
APWS-IC allele as well as a transgene. The relevant transgenic line is listed above each lane. 
5 ug of total brain RNA was loaded on a gel and the blot was probed with a portion of the 
coding region for the Ndn gene. B. RT-PCR was used to investigate Magel2 and MIcrn3 
in the same mice. 5 u,g of RNA was used to program RT-PCR reactions. 





Figure 5-6. Snrpn and snoRNA expression in 380 and 1707 transgenic lines. A. 

Northern analysis was used to show Snrpn expression in mice that had inherited a 
paternal APWS-IC allele as well as a transgene. The relevant transgenes are listed. 5 
u.g of total brain RNA was loaded per lane, and the blot was probed with the 5' portion 
of the Snrpn transcript. B. Northern analysis was performed for snoRNAs. 10 [xg of 
total brain RNA was loaded per lane on a 8% polyacrylamide gel. The blot was probed 
with oligonucleotides complementary to MbII-85, MbII-52, and MbII-13. 



108 









454 Transgenic Females 




30 - 










25 - 








•ams) 


20 - 








■•— ' 


15 ■ 




y(f 




Weigh 


10 - 
5 ■ 
- 






— ♦— wt 

IC deletion 

IC deletion, 454A 

IC deletion, 454B 

— jk — IC deletion, 454D 






3 


1 1 1 

6 9 12 15 

Age (weeks) 


1 1 i 1 

18 21 24 



454 Transgenic Males 

35 n 




3 6 9 12 15 18 21 
Age (weeks) 



Figure 5-7. Transgenes harboring Mkrn3, Magel2, and Ndn do not rescue small 
stature phenotype. Mice inheriting a paternal APWS-IC allele as well as a transgene 
were weighed once every 3 weeks from weaning. Their age in weeks versus weight in 
grams was plotted. 



109 



35 

30 

^25-1 

In 
E 

5 20 



380 Transgenic Females 



5 



15 
10 -I 

5 






— ♦— wt 

IC deletion 

IC deletion, 1707 

IC deletion, 380A 

-*— IC deletion, 380B 

-•—IC deletion, 380D 

— i— wt, 380A 



12 15 
Age (weeks) 



18 



21 



24 



50 n 
45 
40 
^ 35 

V) 

I 30 

o> 

— 25 
£ 

20 

O 

* 15 
10 

5 -j 




380 Transgenic Males 




wt 

IC deletion 
IC deletion, 1707 
IC deletion, 380A 
wt, 380A 



12 15 
Age (weeks) 



18 



21 



24 



Figure 5-8. Transgene 380A rescues small stature phenotype, but other 
transgenes bearing Snrpn do not. Mice inheriting a paternal APWS-IC allele as well 
as a transgene were weighed every 3 weeks beginning at weaning. Their age in weeks 
is plotted versus weight in grams. Wild-type mice harboring the 380A transgene are 
also shown. 



CHAPTER 6 
CONCLUSIONS AND FUTURE DIRECTIONS 

This study of the mechanisms and phenotypes involved with PWS has contributed 

much to the understanding of this disorder, yet there is still much to be learned. 

However, this study demonstrates the multiple uses of a mouse model for a human 

imprinting disorder and how it can be used to dissect the many intricacies associated with 

a disease. 

The maternally expressed genes in the AS/PWS region, UBE3A and ATP IOC, 
seem mysterious since their imprinted expression seems to be limited to specific cell or 
tissue types. UBE3A does not appear to behave as a classical imprinted gene, associated 
with methylation imprints and tightly regulated parent-of-origin expression, and ATP IOC 
has not been well studied to date. The data presented in this dissertation suggest a 
possible model for imprinted expression of murine Ube3a that simplifies the regulation of 
the maternally expressed genes (Fig. 2-4). The data specifically suggest that Ube3a is 
negatively regulated by a paternally expressed transcript antisense to the gene and that 
this transcript is positively regulated by the PWS-IC. This does not provide definitive 
proof that the regulation of Ube3a imprinting occurs via the antisense transcript, 
however. Termination of the antisense transcript to Ube3a could be used to prove this 
hypothesis for gene regulation. Alternatively, demonstration that imprinted expression of 
Ube3a coincides with the presence of the antisense transcript in a particular cell type and 
that in the same cell type the absence of this transcript prevents imprinted Ube3a 



110 



Ill 

expression would provide overwhelming evidence that this model for regulation of 
Ube3a is correct. It would also be interesting to see whether AtplOc is similarly 
regulated. 

The imprinted transgenes identified in this study are not terribly informative with 
respect to the mechanisms of imprinting in the PWS/AS region, but they do create an 
attractive system in which it can be investigated in great detail. The transgenes studied 
demonstrate that correct imprinting of a Snrpn transgene can occur even when they are 
integrated at an ectopic site in the mouse genome. The minimal area necessary to achieve 
appropriate imprinted expression is nearly 170 kb. Some data suggest that this minimal 
region might be as small as 63 kb, however this must be tested in a more rigorous manner 
before it can be believed. These imprinted transgenes also suggest that the AS-IC, which 
has not been functionally identified in mouse, is located within 90 kb of the Snrpn gene. 
These data can be used to make a targeted murine AS-IC mutation. Further experiments 
along these lines include deleting the appropriate regions on the 425D18 BAC to mimic 
the rearrangements seen in transgenic line 215B, with hopes of further narrowing the 
minimal area sufficient to confer proper imprinting, and deleting specific elements within 
the PWS-IC and AS-IC to see if they are required for imprinted expression. A marked 
Snrpn locus has already been created to begin these studies. 

The phenotype studies presented in this dissertation demonstrate the clinical 
relevance of mouse models for PWS. The observation that there are strain dependent 
differences in the phenotype of PWS mice has allowed us to investigate the adult onset 
phenotypes involved with PWS. Since the affected mice can survive, behavioral tests, 
obesity studies, and nutritional studies can now be carried out on these mice. Treatments 



112 

for the disorder can also be tested. These mice also showed a small stature phenotype 
that parallels the growth deficiencies seen in humans with PWS. Perhaps the most 
interesting outcome of this investigation is the identification of modifier genes that may 
affect the phenotype of PWS mice. These genes may identify other genes that are 
involved with PWS, helping to dissect the precise reasons that PWS infants present with 
failure to thrive and poor suckle. Continued studies in this area will focus on the 
identification of this modifying factor using the APWS-IC mouse to map and clone the 
gene(s) responsible. 

Finally, some of the most clinically relevant information has come from the 
transgenic rescue experiments. Two important discoveries are presented in this study. 
First, none of the known paternally expressed genes rescue the failure to thrive phenotype 
in APWS-IC mice. This indicates that there are additional genes involved in the APWS- 
IC phenotype that have not been identified and that Ndn deficiency in mice needs to be 
further studied. Secondly, MbII-85 in concert with Snrpn, MbII-13, and other single copy 
snoRNAs can rescue the small stature phenotype of APWS-IC mice. This second finding 
is particularly important since the most effective therapy for PWS is growth hormone. 
While the rescuing transgene alleviates small stature in APWS-IC mice, it may have other 
effects, only one of which may be activation of the growth hormone pathway. The 
snoRNAs on this transgene may someday provide useful therapy that can ameliorate 
some of the difficulties that humans with PWS face. Additional studies will include 
targeted mutation of the individual snoRNAs and perhaps more specific rescue 
experiments that isolate each snoRNA. The transgenes tested in this study will also be 
provided together to determine their ability to rescue the APWS-IC phenotype. The 



113 



possibility that the rescuing transgene may also reveal an obesity phenotype is intriguing 
and warrants further study. Finally, behavioral phenotypes associated with PWS can also 
be investigated using the established transgenic lines. 



REFERENCES 



1. Barlow, D.P. Gametic imprinting in mammals. Science 270, 1610-1613 (1995). 

2. McGrath, J. & Solter, D. Completion of mouse embryogenesis requires both the 
maternal and paternal genomes. Cell 37, 179-183 (1984). 

3. Barton, S.C., Surani, M.A. & Norris, M.L. Role of paternal and maternal genomes 
in mouse development. Nature 311, 374-376 (1984). 

4. Cattanach, B.M. Parental origin effects in mice. J Embryol Exp Morphol 97, 137- 
50 (1986). 

5. Reik, W. & Allen, N.D. Genomic imprinting. Imprinting with and without 
methylation. Curr-BioU, 145-7 issn: 0960-9822 (1994). 

6. Buiting, K., Saitoh, S., Gross, S., Dittrich, B., Schwartz, S., Nicholls, R.D. & 
Horsthemke, B. Inherited microdeletions in the Angelman and Prader-Willi 
syndromes define an imprinting centre on human chromosome 15. Nature Genet. 
9, 395-400 (1995). 

7. Dittrich, B., Buiting, K., Korn, B., Rickard, S., Buxton, J., Saitoh, S., Nicholls, 
R.N., Poustka, A., Winterpacht, A., Zabel, B. et al. Imprint switching on human 
chromosome 15 may involve alternative transcripts of the SNRPN gene. Nature 
Genet. 14, 163-170(1996). 

8. Fundele, R.H., Surani, M.A. & Allen, N.D. Consequences of genomic imprinting 
for fetal development, in Genomic Imprinting (eds. Reik, W. & Surani, M.A.) 98- 
112 (IRL/Oxford University Press, Oxford, New York, Tokyo, 1997). 

9. Jaenisch, R. & Bird, A. Epigenetic regulation of gene expression: how the 
genome integrates intrinsic and environmental signals. Nature Genet. Suppl. 33, 
245-254 (2003). 

10. Feil, R. & Kelsey, G. Genomic imprinting: A chromatin connection. Am J Hum 
Genet 61, 1213-1219(1997). 

11. Glenn, C.C., Porter, K.A., Jong, M.T., Nicholls, R.D. & Driscoll, D.J. Functional 
imprinting and epigenetic modification of the human SNRPN gene. Hum. Mol. 
Genet. 2, 2001-5 (1993). 

12. Reis, A., Dittrich, B., Greger, V., Buiting, K., Lalande, M., Gillessen Kaesbach, 
G., Anvret, M. & Horsthemke, B. Imprinting mutations suggested by abnormal 
DNA methylation patterns in familial Angelman and Prader-Willi syndromes. 
Am. J. Hum. Genet. 54, 741-7 (1994). 

13. Sutcliffe, J.S., Nakao, M., Christian, S., Orstavik, K.H., Tommerup, N., Ledbetter, 
D.H. & Beaudet, A.L. Deletions of a differentially methylated CpG island at the 
SNRPN gene define a putative imprinting control region. Nature Genet. 8, 52-8 
(1994). 

14. Ledbetter, D.H., Riccardi, V.M., Airhart, S.D., Strobel, R.J., Keenan, B.S. & 
Crawford, J.D. Deletions of chromosome 15 as a cause of the Prader-Willi 
syndrome. N-Engl-J-Med 304, 325-9 issn: 0028-4793 (1981). 



114 



115 



15. Jong, M.T.C., Carey, A.H., Stewart, C.L., Rinchik, E.M., Glenn, C.C., Driscoll, 
D.J. & Nicholls, R.D. The ZNF127 gene encodes a novel C3H4 zinc-finger 
protein and its expression is regulated by genomic imprinting. Am. J. Hum. Genet. 
53, A697 (1993). 

16. Jay, P., Rougeulle, C, Massacrier, A., Moncla, A., Mattei, M.-G., Malzac, P., 
Roeckel, N., Taviaux, S., Lefranc, J.-L.B., Cau, P. et al. The human necdin gene, 
NDN, is maternally imprinted and located in the Prader-Willi syndrome 
chromosomal region. Nature Genet. 17, 357-361 (1997). 

17. Leff, S.E., Brannan, C.I., Reed, M.L., Ozcelik, T., Francke, U., Copeland, N.G. & 
Jenkins, NA. Maternal imprinting of the mouse Snrpn gene and conserved 
linkage homology with the human Prader-Willi syndrome region. Nature Genet. 
2, 259-64 (1992). 

18. Wevrick, R., Kerns, J.A. & Francke, U. Identification of a novel paternally 
expressed gene in the Prader-Willi syndrome region. Hum. Mol. Genet. 3, 1877- 
82 (1994). 

19. Rougeulle, C., Glatt, H. & Lalande, M. The Angelman syndrome candidate gene, 
UBE3AJE6-AP, is imprinted in the brain. Nature Genet. 17, 14-15 (1997). 

20. Saitoh, S., Buiting, K., Rogan, P.K., Buxton, J.L., Driscoll, D.J., Arnemann, J., 
Konig, R., Malcolm, S., Horsthemke, B. & Nicholls, R.D. Minimal definition of 
the imprinting center and fixation of chromosome 15ql l-ql3 epigenotype by 
imprinting mutations. Proc. Natl. Acad. Sci. USA 93, 7811-7815 (1996). 

21. Cattanach, B.M., Barr, J.A., Evans, E.P., Burtenshaw, M., Beechey, C.V., Leff, 
S.E., Brannan, C.I., Copeland, N.G., Jenkins, N.A. & Jones, J. A candidate mouse 
model for Prader-Willi syndrome which shows an absence of Snrpn expression. 
Nature Genet. 2, 270-4 (1992). 

22. Yang, T., Adamson, T.E., Resnick, J.L., Leff, S., Wevrick, R., Francke, U., 
Jenkins, N.A., Copeland, N.G. & Brannan, C.I. A mouse model for Prader-Willi 
syndrome imprinting-centre mutations. Nature Genet. 19, 25-31 (1998). 

23. Holm, V.A., Cassidy, S.B., Butler, M.G., Hanchett, J.M., Greenswag, L.R. 
Whitman, B.Y. and Greenberg, F. Prader-Willi syndrome: Consensus Diagnostic 
Criteria. Pediatrics 91, 398-402 (1993). 

24. Clayton-Smith, J. & Pembrey, M.E. Angelman syndrome. Med. Genet. 29, 412- 
415 (1992). 

25. Cassidy, S.B. & Ledbetter, D.H. Prader-Willi syndrome. Neurol-Clin 7, 37-54 
issn: 0733-8619 (1989). 

26. Nicholls, R.D., Saitoh, S. & Horsthemke, B. Imprinting in Prader-Willi and 
Angelman syndromes. Trends Genet. 14, 194-200 (1998). 

27. Butler, M.G. & Palmer, C.G. Parental origin of chromosome 15 deletion in 
Prader-Willi syndrome [letter]. Lancet 1, 1285-6 issn: 0140-6736 (1983). 

28. Nicholls, R.D., Knoll, J.H., Butler, M.G., Karam, S. & Lalande, M. Genetic 
imprinting suggested by maternal heterodisomy in nondeletion Prader-Willi 
syndrome. Nature 342, 281-5 (1989). 

29. Magenis, R.E., Brown, M.G., Lacy, D.A., Budden, S. & LaFranchi, S. Is 
Angelman syndrome an alternate result of del(15)(ql lql3)? Am-J-Med-Genet 28, 
829-38 issn: 0148-7299 (1987). 



116 



30. Knoll, J.H., Nicholls, R.D. & Lalande, M. On the parental origin of the deletion in 
Angelman syndrome. Hum-Genet 83, 205-7 (1989). 

3 1 . Malcolm, S., Clayton Smith, J., Nichols, M., Robb, S., Webb, T., Armour, J.A., 
Jeffreys, A.J. & Pembrey, M.E. Uniparental paternal disomy in Angelman's 
syndrome. Lancet 337, 694-7 (1991). 

32. Knoll, J.H., Glatt, K.A., Nicholls, R.D., Malcolm, S. & Lalande, M. Chromosome 
15 uniparental disomy is not frequent in Angelman syndrome. Am- J -Hum-Genet 
48, 16-21 issn: 0002-9297 (1991). 

33. Cox, G.F., Burger, J., Lip, V., Mau, U.A., Sperling, K., Wu, B.L. & Horsthemke, 
B. Intracytoplasmic sperm injection may increase the risk of imprinting defects. 
Am J Hum Genet 71, 162-164 (2002). 

34. Kishino, T., Lalande, M. & Wagstaff, J. UBE3AJE6-AP mutations cause 
Angelman syndrome. Nature Genet. 15, 70-73 (1997). 

35. Matsuura, T., Sutcliffe, J.S., Fang, P., Galjaard, R.-J., Jiang, Y.-H., Benton, C.S., 
Rommens, J.M. & Beaudet, A.L. De novo truncating mutations in E6-AP 
ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nature Genet. 15, 
74-77(1997). 

36. Meguro M, K.A., Mitsuya K, Nakao M, Kondo I, Saitoh S, Oshimura, M. A novel 
maternally expressed gene, ATP 10C encodes a putative aminophospholipid 
translocase associated with Angelman syndrome. Nature Genet. 28, 19-20 (2001). 

37. Ohta, T., Gray, T.A., Rogan, P.K., Buiting, K., Gabriel, J.M., Saitoh, S., 
Muralidhar, B., Bilienska, B., Krajewska-Walasek, M., Driscoll, D.J. et al. 
Imprinting-mutation mechanisms in Prader-Willi syndrome. Am. J. Hum. Genet. 
64, 397-413 (1999). 

38. Ohta, T., Buiting, K., Kokkonen, H., McCandless, S., Heeger, S., Leisti, H., 
Driscoll, D.J., Cassidy, S.B., Horsthemke, B. & Nicholls, R.D. Molecular 
mechanism of Angelman syndrome in two large families involves an imprinting 
mutation. Am. J. Hum. Genet. 64, 385-396 (1999). 

39. Ozcelik, T., Leff, S., Robinson, W., Donlon, T., Lalande, M., Sanjines, E., 
Schinzel, A. & Franke, U. Small nuclear ribonucleoprotein polypeptide N 
(SNRPN), an expressed gene in the Prader-Willi syndrome critical region. Nature 
Genet. 2, 265-269 (1992). 

40. MacDonald, H.R. & Wevrick, R. The necdin gene is deleted in Prader-Willi 
syndrome and is imprinted in human and mouse. Hum. Mol. Genet. 6, 1873-1878 
(1997). 

41. Rougeulle, C, Cardoso, C, Fontes, M., Colleaux, L. & Lalande, M. An imprinted 
antisense RNA overlaps UBE3A and a second maternally expressed transcript. 
Nature Genet. 19, 15-16 (1998). 

42. Boccaccio, I., Glatt-Deeley, H., Watrin, F., Roeckel, N., Lalande, M. & 
Muscatelli, F. The human MAGEL2 gene and its mouse homologue are paternally 
expressed and mapped to the Prader-Willi region. Hum. Molec. Genet. 8, 2497- 
2505 (1999). 

43. Cavaille, J., Buiting, K., Kiefman, M., M., L., C.I., B., Horsthemke, B., 
Bachellerie, J.-P., Brosius, J. & Huttenhoffer, A. Identification of brain-specific 
and imprinted small nucleolar RNA genes exhibiting an unusual genomic 
organization. Proc. Nat. Acad. Sci. USA 97, 1431 1-6 (2000). 



117 



44. de los Santos, T., Schweizer, J., Rees, C.A. & Francke, U. Small evolutionarily 
conserved RNA, resembling C/D box small nucleolar RNA, is transcribed from 
the PWCR1, a novel imprinted gene in the Prader-Willi deleted region, which is 
highly expressed in the brain. Am. J. Hum. Genet. 67, 1067-1082 (2000). 

45. Knoll, J.H., Nicholls, R.D., Magenis, R.E., Graham, J.M., Jr., Lalande, M. & Latt, 
S.A. Angelman and Prader-Willi syndromes share a common chromosome 15 
deletion but differ in parental origin of the deletion. Am. J. Med. Genet. 32, 285- 
90 (1989). 

46. Jong, M.T., Carey, A.H., Caldwell, K.A., Lau, M.H., Handel, M.A., Driscoll, 
D.J., Stewart, C.L., Rinchik, E.M. & Nicholls, R.D. Imprinting of a RING zinc- 
finger encoding gene in the mouse chromosome region homologous to the Prader- 
Willi syndrome genetic region. Hum. Mol. Genet. 8, 795-803 (1999). 

47. Wevrick, R. & Francke, U. An imprinted mouse transcript homologous to the 
human Imprinted in Prader-Willi syndrome (IPW) gene. Hum. Molec. Genet. 6, 
325-332 (1997). 

48. Albrecht, U., Sutcliffe, J.S., Cattanach, B.M., Beechy, C.V., Armstrong, D., 
Eichele, G. & Beaudet, A.L. Imprinted expression if the murine Angelman 
syndrome gene, Ube3a, in hippocampal and Purkinje neurons. Nature Genet. 17, 
75-78 (1997). 

49. Mascari, M.J., Gottlieb, W., Rogan, P.K., Butler, M.G., Waller, D.A., Armour, 
J. A., Jeffreys, A.J., Ladda, R.L. & Nicholls, R.D. The frequency of uniparental 
disomy in Prader-Willi syndrome. Implications for molecular diagnosis. N. Engl. 
J. Med. 326, 1599-607 (1992). 

50. Donlon, T.A. Similar molecular deletions on chromosome 15ql 1.2 are 
encountered in both the Prader-Willi and Angelman syndromes. Hum-Genet 80, 
322-328 (1988). 

5 1 . Nicholls, R.D. Genomic imprinting and uniparental disomy in Angelman and 
Prader-Willi syndromes: a review. Am-J-Med-Genet 46, 16-25 issn: 0148-7299 
(1993). 

52. Chan, C.T., Clayton Smith, J., Cheng, X.J., Buxton, J., Webb, T., Pembrey, M.E. 
& Malcolm, S. Molecular mechanisms in Angelman syndrome: a survey of 93 
patients. J-Med-Genet 30, 895-902 issn: 0022-2593 (1993). 

53. Vu, T.H. & Hoffman, A.R. Imprinting of the Angelman syndrome gene, UBE3A, 
is restricted to the brain. Nature Genet. 17, 12-13 (1997). 

54. Mowery-Rushton, P.A., Driscoll, D.J., Nicholls, R.D., Locker, J. & Surti, U. 
DNA methylation patterns in human tissues of uniparental origin using a zinc- 
finger gene (ZNF127) from the Angelman/Prader- Willi region. Am J Med Genet 
61, 140-146 (1996). 

55. Bielinska, B., Blaydes, S.M., Buiting, K., Yang, T., Krajewska-Walasek, M., 
Horsthemke, B. & Brannan, C.I. De novo deletions of SNRPN exon 1 in early 
human and mouse embryos result in a paternal to maternal imprint switch. Nature 
Genet. 25, 74-8 (2000). 

56. Wakeland, E.K., Morel, L., Achey, K., Yui, M. & Longmate, J. Speed congenics: 
A classic technique in the fast lane (relatively speaking). Immunology Today 18, 
472-477 (1997). 



118 



57. Brannan, C.I. & Bartolomei, M.S. Mechanisms of genomic imprinting. Curr Opin 
Genet Dev 9, 164-70(1999). 

58. Schedl, A., Montoliu, L., Kelsey, G. & Schutz, G. A yeast artificial chromosome 
covering the tyrosinase gene confers copy number-dependent expression in 
transgenic mice. Nature 362, 258-261 (1993). 

59. Montoliu, L., Umland, T. & Schutz, G. A locus control region at -12 kb of the 
tyrosinase gene. EMBO 15, 6026-6034 (1996). 

60. Fraser, P., Pruzina, S. & Grosveld, F. Each hypersensitive site of the human |3- 
globin locus control region confers a different developmental pattern of 
expression to the globin genes. Genes & Dev. 7, 106-1 13 (1993). 

6 1 . Peterson KR, C.C., Huxley C, Josephson BM, Haugen HS, Furukawa T, 
Stamatoyannopoulos G. Transgenic mice containing a 248-kb yeast artificial 
chromosome carrying the human (3-globin locus display proper developmental 
control of human globin genes. Proc. Nat. Acad. Sci. USA 90, 7593-7597 (1993). 

62. Bungert, J., Dave, U., Lim, K.-C, Shavit, J.A., Liu, Q. & Engel, J.D. Synergistic 
regulation of human b-globin gene switching by locus control region elements 
HS3 and HS4. Genes & Dev. 9, 3083-3096 (1995). 

63. Wutz, A., Smrzka, O.W., Schweifer, N., Schellander, K., Wagner, E.F. & Barlow, 
D.P. Imprinted expression of the Igf2r gene depends on an intronic CpG island. 
Nature 389, 745-749 (1997). 

64. Bartolomei, M.S., Webber, A.L., Brunkow, M.E. & Tilghman, S.M. Epigenetic 
mechanisms underlying the imprinting of the mouse H19 gene. Genes Dev. 7, 
1663-73 (1993). 

65. Pfeifer, K., Leighton, P.A. & Tilghman, S.M. The structural H19 gene is required 
for transgene imprinting. Proc. Natl. Acad. Sci. USA 93, 13876-13883 (1996). 

66. Hark AT, S.C., Katz DJ, Ingram RS Levorse JM, Tilghman SM. CTCF mediates 
methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature 
405, 486-489 (2000). 

67. Cranston MJ, S.T., Elson DA, Bartolomei MS. Elucidation of the minimal 
sequence required to imprint H19 transgenes. Genomics 73, 98-107 (2001). 

68. Ainscough, J.F.-X., Koide, T., Tada, M., Barton, S. & Surani, A. Imprinting of 
Igf2 and H19 from a 130 kb YAC transgene. Development 124, 3621-3632 
(1997). 

69. Bell AC, F.G. Methylation of a CTCF-dependent boundary controls imprinted 
expression of the Igf2 gene. Nature 405, 482-485 (2000). 

70. Blaydes, S.M., Elmore, M., Yang, T. & Brannan, C.I. Analysis of murine Snrpn 
and human SNRPN gene imprinting in transgenic mice. Mamm. Genome 10, 549- 
555 (1999). 

71. Yang, X.W., Model, P. & Heintz, N. Homologous recombination based 
modification in Esherichia coli and germline transmission in transgenic mice of a 
bacterial artificial chromosome. Nature Biotechnology 15, 859-865 (1997). 

72. Zhang Y, B.F., Muyrers JPP, Stewart AF. A new logic for DNA engineering 
using recombination in Escherichia coli. Nature Genet. 20, 123-128 (1998). 

73. Lee EC, Y.D., de Velasco JM, Tessarollo L, Swing DA, Court DL, Jenkins NA, 
Copeland NG. A highly efficient Escherichia co/t'-based chromosome engineering 



119 



system adapted for recombinogenic targeting and subcloning of B AC DNA. 
Genomics 73, 56-65 (2001). 

74. Holm, V.A., Cassidy, S.B., Butler, M.G., Hanchett, J.M., Greenswag, L.R., 
Whitman, B.Y. & Greenberg, F. Prader-Willi syndrome: consensus diagnostic 
criteria. Pediatrics 91, 398-402 (1993). 

75. Nicholls, R.D., Knoll, J.H., Glatt, K., Hersh, J.H., Brewster, T.D., Graham, J.M., 
Jr., Wurster Hill, D., Wharton, R. & Latt, S.A. Restriction fragment length 
polymorphisms within proximal 15q and their use in molecular cytogenetics and 
the Prader-Willi syndrome. Am. J. Med. Genet. 33, 66-77 (1989). 

76. Robinson, W.P., Bottani, A., Xie, Y.G., Balakrishman, J., Binkert, F., Machler, 
M., Prader, A. & Schinzel, A. Molecular, cytogenetic, and clinical investigations 
of Prader-Willi syndrome patients. Am. J. Hum. Genet. 49, 1219-34 (1991). 

77. Magenis, R.E., Toth Fejel, S., Allen, L.J., Black, M., Brown, M.G., Budden, S., 
Cohen, R., Friedman, J.M., Kalousek, D., Zonana, J. et al. Comparison of the 15q 
deletions in Prader-Willi and Angelman syndromes: specific regions, extent of 
deletions, parental origin, and clinical consequences. Am. J. Med. Genet. 35, 333- 
49(1990). 

78. Williams, C.A., Zori, R.T., Stone, J.W., Gray, B.A., Cantu, E.S. & Ostrer, H. 
Maternal origin of 15ql 1-13 deletions in Angelman syndrome suggests a role for 
genomic imprinting. Am. J. Med. Genet. 35, 350-3 (1990). 

79. Nicholls, R.D., Pai, G.S., Gottlieb, W. & Cantu, E.S. Paternal uniparental disomy 
of chromosome 15 in a child with Angelman syndrome. Ann. Neurol. 32, 512-518 
(1992). 

80. Nicholls, R.D. New insights reveal complex mechanisms involved in genomic 
imprinting. Am. J. Hum. Genet. 54, 733-40 (1994). 

8 1 . Jong, M.T., Gray, T. A., Ji, Y., Glenn, C.C., Saitoh, S., Driscoll, D.J. & Nicholls, 
R.D. A novel imprinted gene, encoding a RING zinc-finger protein, and 
overlapping antisense transcript in the Prader-Willi syndrome critical region. 
Hum. Mol. Genet. 8, 783-793 (1999). 

82. Reed, M.L. & Leff, S.E. Maternal imprinting of human SNRPN, a gene deleted in 
Prader-Willi syndrome. Nature Genet. 6, 163-7 (1994). 

83. Runte, M., Huttenhofer, A., Gross, S., Kiefmann, M., Horsthemke, B. & Buiting, 
K. The IC-SNURF-SNRPN transcript serves as a host for multiple small nucleolar 
RNA species and as an antisense RNA for Ube3a. Hum. Mol. Genet. 10, 2687- 
2700 (2001). 

84. Lee, S., Kozlov, S., Hernandez, L. Chamberlain, S.J., Brannan, C.I., Stewart, 
C.L., and Wevrick, R. Expression and imprinting of MAGEL2 suggest a role in 
Prader-Willi syndrome and the homologous murine imprinting phenotype. Hum. 
Mol. Genet. 9, 1813-1819 (2000). 

85. Sun, Y., Nicholls, R.N., Bulter, M.G., Saitoh, S., Hainline, B.E. & Palmer, C.G. 
Breakage in the Snrpn locus in a balanced 46,XY,t(15;19) Prader-Willi syndrome 
patient. Hum. Molec. Genet. 5, 517-524 (1996). 

86. Muralidhar, B., Marney, A. & Butler, M.G. Analysis of imprinted genes in 
subjects with Prader-Willi syndrome and chromosome 15 abnormalities. Genet. 
Med. 1, 141-145 (1999). 



120 



87. Bressler, J., Tsai, T., Ramirez, M., Armstrong, D. & Beaudet, A. The SNPRN 
promoter is not required for genomic imprinting of the Prader-Willi/ Angelman 
domain in mice. Nature Genet. 28, 232-240 (2001). 

88. Budarf ML, E.B. Progress in the autosomal segmental aneusomy dyndromes 
(SASs): single or multi-locus disorders? Hum. Mol. Genet. 6, 1657-1665 (1997). 

89. Tsai TF, A.D., Beaudet, AL. Atecdm-deficient mice do not show lethality or the 
obesity and infertility of Prader-Willi syndrome. Nature Genet. 22, 15-16 (1999). 

90. Gerard M, H.L., Wevrick R, Stewart CL. Disruption of the mouse necdin gene 
results in early post-natal lethality. Nature Genet. 23, 199-202 (1999). 

9 1 . Tsai TF, J.Y., Bressler J, Armstrong D, Beaudet, AL. Paternal deletion from 
Snrpn to Ube3a in the mouse causes hypotonia, growth retardation, and partial 
lethality and provides evidence for a gene contributing to Prader-Willi syndrome. 
Hum. Mol. Genet. 8, 1357-1364 (1999). 

92. Muscatelli, F., Abrous, D.N., Massacrier, A., Boccaccio, I., Le Moal, M., Cau, P. 
& Cremer, H. Disruption of the mouse Necdin gene results in hypothalamic and 
behavioral alterations reminiscent of the human Prader-Willi syndrome. Hum. 
Mol. Genet. 9, 3101-3110 (2000). 



BIOGRAPHICAL SKETCH 

Stormy Jo Chamberlain was born in Moscow, Idaho and grew up on a very nice 
tractor in the middle of nowhere near Casper, Wyoming. She graduated from Natrona 
County High School in Casper, Wyoming in 1993. She then received her BA. degree in 
Molecular Biology from Princeton University in 1997. Stormy entered the 
Interdisciplinary Program in Biomedical Sciences at the University of Florida in 1997. 



121 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 



James Resnick, Chair 

Associate Professor of Molecular Genetics and 
Microbiology 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Daniel J. Driscoll 

Professor of Molecular Genetics and Microbiology 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

1 




Richard Moyer, 

Professor of Molecular Genetics and Microbiology 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Jorg Bungert 
Assistant Professor of Biochemisty and Molecular 
Biology 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Laurence Morel 

Assistant Professor of Pathology, Immunology, and 
Laboratory Medicine 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 



Mark Lewis 

Professor of Neuroscience 



This dissertation was submitted to the Graduate Faculty of the College of 
Medicine and to the Graduate School and was accepted as partial fulfillment of the 
requirements for the degree of Doctor of Philosophy. 

May 2003 

e. L^JAi — 

Dean, College of Medicine 




Dean, Graduate School